Compare commits

..

4 Commits

Author SHA1 Message Date
henrygd
1275af956b updates 2025-11-24 16:57:06 -05:00
henrygd
bf36015bd9 updates 2025-11-24 16:40:18 -05:00
henrygd
56807dc5e4 updates 2025-11-21 17:49:17 -05:00
henrygd
56a9915b43 quiet hours progress 2025-11-21 17:09:42 -05:00
203 changed files with 8494 additions and 22014 deletions

2
.github/CODEOWNERS vendored
View File

@@ -1,2 +0,0 @@
# Everything needs to be reviewed by Hank
* @henrygd

View File

@@ -1,19 +0,0 @@
body:
- type: dropdown
id: component
attributes:
label: Component
description: Which part of Beszel is this about?
options:
- Hub
- Agent
- Hub & Agent
default: 0
validations:
required: true
- type: textarea
attributes:
label: Description
description: Please describe in detail what you want to share.
validations:
required: true

View File

@@ -1,54 +1,19 @@
body:
- type: checkboxes
id: terms
- type: markdown
attributes:
label: Welcome!
description: |
Thank you for reaching out to the Beszel community for support! To help us assist you better, please make sure to review the following points before submitting your request:
value: |
### Before opening a discussion:
Please note:
- For translation-related issues or requests, please use the [Crowdin project](https://crowdin.com/project/beszel).
**- Please do not submit support reqeusts that are specific to ZFS. We plan to add integration with ZFS utilities in the near future.**
options:
- label: I have read the [Documentation](https://beszel.dev/guide/getting-started)
required: true
- label: I have checked the [Common Issues Guide](https://beszel.dev/guide/common-issues) and my problem was not mentioned there.
required: true
- label: I have searched open and closed issues and discussions and my problem was not mentioned before.
required: true
- label: I have verified I am using the latest version available. You can check the latest release [here](https://github.com/henrygd/beszel/releases).
required: true
- type: dropdown
id: component
attributes:
label: Component
description: Which part of Beszel is this about?
options:
- Hub
- Agent
- Hub & Agent
default: 0
validations:
required: true
- Check the [common issues guide](https://beszel.dev/guide/common-issues).
- Search existing [issues](https://github.com/henrygd/beszel/issues) and [discussions](https://github.com/henrygd/beszel/discussions) (including closed).
- type: textarea
id: description
attributes:
label: Problem Description
description: |
How to write a good bug report?
- Respect the issue template as much as possible.
- The title should be short and descriptive.
- Explain the conditions which led you to report this issue: the context.
- The context should lead to something, a problem that youre facing.
- Remain clear and concise.
- Format your messages to help the reader focus on what matters and understand the structure of your message, use [Markdown syntax](https://help.github.com/articles/github-flavored-markdown)
label: Description
description: A clear and concise description of the issue or question. If applicable, add screenshots to help explain your problem.
validations:
required: true
- type: input
id: system
attributes:
@@ -56,15 +21,13 @@ body:
placeholder: linux/amd64 (agent), freebsd/arm64 (hub)
validations:
required: true
# - type: input
# id: version
# attributes:
# label: Beszel version
# placeholder: 0.9.1
# validations:
# required: true
- type: input
id: version
attributes:
label: Beszel version
placeholder: 0.9.1
validations:
required: true
- type: dropdown
id: install-method
attributes:
@@ -78,21 +41,18 @@ body:
- Other (please describe above)
validations:
required: true
- type: textarea
id: config
attributes:
label: Configuration
description: Please provide any relevant service configuration
render: yaml
- type: textarea
id: hub-logs
attributes:
label: Hub Logs
description: Check the logs page in PocketBase (`/_/#/logs`) for relevant errors (copy JSON).
render: json
- type: textarea
id: agent-logs
attributes:

View File

@@ -1,30 +1,8 @@
name: 🐛 Bug report
description: Use this template to report a bug or issue.
description: Report a new bug or issue.
title: '[Bug]: '
labels: ['bug']
labels: ['bug', "needs confirmation"]
body:
- type: checkboxes
attributes:
label: Welcome!
description: |
The issue tracker is for reporting bugs and feature requests only. For end-user related support questions, please use the **[GitHub Discussions](https://github.com/henrygd/beszel/discussions/new?category=support)** instead
Please note:
- For translation-related issues or requests, please use the [Crowdin project](https://crowdin.com/project/beszel).
- To request a change or feature, use the [feature request form](https://github.com/henrygd/beszel/issues/new?template=feature_request.yml).
- Any issues that can be resolved by consulting the documentation or by reviewing existing open or closed issues will be closed.
**- Please do not submit bugs that are specific to ZFS. We plan to add integration with ZFS utilities in the near future.**
options:
- label: I have read the [Documentation](https://beszel.dev/guide/getting-started)
required: true
- label: I have checked the [Common Issues Guide](https://beszel.dev/guide/common-issues) and my problem was not mentioned there.
required: true
- label: I have searched open and closed issues and my problem was not mentioned before.
required: true
- label: I have verified I am using the latest version available. You can check the latest release [here](https://github.com/henrygd/beszel/releases).
required: true
- type: dropdown
id: component
attributes:
@@ -34,53 +12,81 @@ body:
- Hub
- Agent
- Hub & Agent
default: 0
validations:
required: true
- type: markdown
attributes:
value: |
### Thanks for taking the time to fill out this bug report!
- For more general support, please [start a support thread](https://github.com/henrygd/beszel/discussions/new?category=support).
- To request a change or feature, use the [feature request form](https://github.com/henrygd/beszel/issues/new?template=feature_request.yml).
- Please do not submit bugs that are specific to ZFS. We plan to add integration with ZFS utilities in the near future.
### Before submitting a bug report:
- Check the [common issues guide](https://beszel.dev/guide/common-issues).
- Search existing [issues](https://github.com/henrygd/beszel/issues) and [discussions](https://github.com/henrygd/beszel/discussions) (including closed).
- type: textarea
id: description
attributes:
label: Problem Description
description: |
How to write a good bug report?
- Respect the issue template as much as possible.
- The title should be short and descriptive.
- Explain the conditions which led you to report this issue: the context.
- The context should lead to something, a problem that youre facing.
- Remain clear and concise.
- Format your messages to help the reader focus on what matters and understand the structure of your message, use [Markdown syntax](https://help.github.com/articles/github-flavored-markdown)
label: Description
description: Explain the issue you experienced clearly and concisely.
placeholder: I went to the coffee pot and it was empty.
validations:
required: true
- type: textarea
id: expected-behavior
attributes:
label: Expected Behavior
description: |
In a perfect world, what should have happened?
**Important:** Be specific. Vague descriptions like "it should work" are not helpful.
description: In a perfect world, what should have happened?
placeholder: When I got to the coffee pot, it should have been full.
validations:
required: true
- type: textarea
id: steps-to-reproduce
attributes:
label: Steps to Reproduce
description: |
Provide detailed, numbered steps that someone else can follow to reproduce the issue.
**Important:** Vague descriptions like "it doesn't work" or "it's broken" will result in the issue being closed.
Include specific actions, URLs, button clicks, and any relevant data or configuration.
description: Describe how to reproduce the issue in repeatable steps.
placeholder: |
1. Go to the coffee pot.
2. Make more coffee.
3. Pour it into a cup.
4. Observe that the cup is empty instead of full.
validations:
required: true
- type: dropdown
id: category
attributes:
label: Category
description: Which category does this relate to most?
options:
- Metrics
- Charts & Visualization
- Settings & Configuration
- Notifications & Alerts
- Authentication
- Installation
- Performance
- UI / UX
- Other
validations:
required: true
- type: dropdown
id: metrics
attributes:
label: Affected Metrics
description: If applicable, which specific metric does this relate to most?
options:
- CPU
- Memory
- Storage
- Network
- Containers
- GPU
- Sensors
- Other
validations:
required: true
- type: input
id: system
attributes:
@@ -88,7 +94,6 @@ body:
placeholder: linux/amd64 (agent), freebsd/arm64 (hub)
validations:
required: true
- type: input
id: version
attributes:
@@ -96,7 +101,6 @@ body:
placeholder: 0.9.1
validations:
required: true
- type: dropdown
id: install-method
attributes:
@@ -110,21 +114,18 @@ body:
- Other (please describe above)
validations:
required: true
- type: textarea
id: config
attributes:
label: Configuration
description: Please provide any relevant service configuration
render: yaml
- type: textarea
id: hub-logs
attributes:
label: Hub Logs
description: Check the logs page in PocketBase (`/_/#/logs`) for relevant errors (copy JSON).
render: json
- type: textarea
id: agent-logs
attributes:

View File

@@ -1,8 +1,5 @@
blank_issues_enabled: false
contact_links:
- name: 🗣️ Translations
url: https://crowdin.com/project/beszel
about: Please report translation issues and request new translations here.
- name: 💬 Support and questions
url: https://github.com/henrygd/beszel/discussions
about: Ask and answer questions here.

View File

@@ -1,25 +1,8 @@
name: 🚀 Feature request
description: Request a new feature or change.
title: "[Feature]: "
labels: ["enhancement"]
labels: ["enhancement", "needs review"]
body:
- type: checkboxes
attributes:
label: Welcome!
description: |
The issue tracker is for reporting bugs and feature requests only. For end-user related support questions, please use the **[GitHub Discussions](https://github.com/henrygd/beszel/discussions)** instead
Please note:
- For **Bug reports**, use the [Bug Form](https://github.com/henrygd/beszel/issues/new?template=bug_report.yml).
- Any requests for new translations should be requested within the [crowdin project](https://crowdin.com/project/beszel).
- Create one issue per feature request. This helps us keep track of requests and prioritize them accordingly.
options:
- label: I have searched open and closed feature requests to make sure this or similar feature request does not already exist.
required: true
- label: This is a feature request, not a bug report or support question.
required: true
- type: dropdown
id: component
attributes:
@@ -29,29 +12,65 @@ body:
- Hub
- Agent
- Hub & Agent
default: 0
validations:
required: true
- type: textarea
id: description
- type: markdown
attributes:
label: Description
description: |
Describe the solution or feature you'd like. Explain what problem this solves or what value it adds.
**Important:** Be specific and detailed. Vague requests like "make it better" will be closed.
placeholder: |
Example:
- What is the feature?
- What problem does it solve?
- How should it work?
value: Before submitting, please search existing [issues](https://github.com/henrygd/beszel/issues) and [discussions](https://github.com/henrygd/beszel/discussions) (including closed).
- type: textarea
attributes:
label: Describe the feature you would like to see
validations:
required: true
- type: textarea
id: motivation
attributes:
label: Motivation / Use Case
description: Why do you want this feature? What problem does it solve?
validations:
required: true
- type: textarea
attributes:
label: Describe how you would like to see this feature implemented
validations:
required: true
- type: textarea
id: logs
attributes:
label: Screenshots
description: Please attach any relevant screenshots, such as images from your current solution or similar implementations.
validations:
required: false
- type: dropdown
id: category
attributes:
label: Category
description: Which category does this relate to most?
options:
- Metrics
- Charts & Visualization
- Settings & Configuration
- Notifications & Alerts
- Authentication
- Installation
- Performance
- UI / UX
- Other
validations:
required: true
- type: dropdown
id: metrics
attributes:
label: Affected Metrics
description: If applicable, which specific metric does this relate to most?
options:
- CPU
- Memory
- Storage
- Network
- Containers
- GPU
- Sensors
- Other
validations:
required: true

View File

@@ -6,7 +6,6 @@ on:
workflow_dispatch:
permissions:
actions: write
issues: write
pull-requests: write
@@ -49,16 +48,11 @@ jobs:
# Action can not skip PRs, set it to 100 years to cover it.
days-before-pr-stale: 36524
# Max issues to process before early exit. Next run resumes from cache. GH API limit: 5000.
operations-per-run: 1500
# Labels
stale-issue-label: 'stale'
remove-stale-when-updated: true
any-of-labels: 'awaiting-requester'
exempt-issue-labels: 'enhancement'
only-issue-labels: 'awaiting-requester'
# Exemptions
exempt-assignees: true
exempt-milestones: true
exempt-milestones: true

View File

@@ -0,0 +1,82 @@
name: Label issues from dropdowns
on:
issues:
types: [opened]
jobs:
label_from_dropdown:
runs-on: ubuntu-latest
permissions:
issues: write
steps:
- name: Apply labels based on dropdown choices
uses: actions/github-script@v7
with:
script: |
const issueNumber = context.issue.number;
const owner = context.repo.owner;
const repo = context.repo.repo;
// Get the issue body
const body = context.payload.issue.body;
// Helper to find dropdown value in the body (assuming markdown format)
function extractSectionValue(heading) {
const regex = new RegExp(`### ${heading}\\s+([\\s\\S]*?)(?:\\n###|$)`, 'i');
const match = body.match(regex);
if (match) {
// Get the first non-empty line after the heading
const lines = match[1].split('\n').map(l => l.trim()).filter(Boolean);
return lines[0] || null;
}
return null;
}
// Extract dropdown selections
const category = extractSectionValue('Category');
const metrics = extractSectionValue('Affected Metrics');
const component = extractSectionValue('Component');
// Build labels to add
let labelsToAdd = [];
if (category) labelsToAdd.push(category);
if (metrics) labelsToAdd.push(metrics);
if (component) labelsToAdd.push(component);
// Get existing labels in the repo
const { data: existingLabels } = await github.rest.issues.listLabelsForRepo({
owner,
repo,
per_page: 100
});
const existingLabelNames = existingLabels.map(l => l.name);
// Find labels that need to be created
const labelsToCreate = labelsToAdd.filter(label => !existingLabelNames.includes(label));
// Create missing labels (with a default color)
for (const label of labelsToCreate) {
try {
await github.rest.issues.createLabel({
owner,
repo,
name: label,
color: 'ededed' // light gray, you can pick any hex color
});
} catch (e) {
// Ignore if label already exists (race condition), otherwise rethrow
if (!e || e.status !== 422) throw e;
}
}
// Now apply all labels (they all exist now)
if (labelsToAdd.length > 0) {
await github.rest.issues.addLabels({
owner,
repo,
issue_number: issueNumber,
labels: labelsToAdd
});
}

1
.gitignore vendored
View File

@@ -10,7 +10,6 @@ dist
*.exe
internal/cmd/hub/hub
internal/cmd/agent/agent
agent.test
node_modules
build
*timestamp*

View File

@@ -16,21 +16,10 @@ builds:
goos:
- linux
- darwin
- windows
- freebsd
goarch:
- amd64
- arm64
- arm
ignore:
- goos: windows
goarch: arm64
- goos: windows
goarch: arm
- goos: freebsd
goarch: arm64
- goos: freebsd
goarch: arm
- id: beszel-agent
binary: beszel-agent
@@ -76,18 +65,6 @@ builds:
- goos: windows
goarch: riscv64
- id: beszel-agent-linux-amd64-glibc
binary: beszel-agent
main: internal/cmd/agent/agent.go
env:
- CGO_ENABLED=0
flags:
- -tags=glibc
goos:
- linux
goarch:
- amd64
archives:
- id: beszel-agent
formats: [tar.gz]
@@ -101,15 +78,6 @@ archives:
- goos: windows
formats: [zip]
- id: beszel-agent-linux-amd64-glibc
formats: [tar.gz]
ids:
- beszel-agent-linux-amd64-glibc
name_template: >-
{{ .Binary }}_
{{- .Os }}_
{{- .Arch }}_glibc
- id: beszel
formats: [tar.gz]
ids:
@@ -118,9 +86,6 @@ archives:
{{ .Binary }}_
{{- .Os }}_
{{- .Arch }}
format_overrides:
- goos: windows
formats: [zip]
nfpms:
- id: beszel-agent
@@ -158,7 +123,9 @@ nfpms:
- debconf
scripts:
templates: ./supplemental/debian/templates
config: ./supplemental/debian/config.sh
# Currently broken due to a bug in goreleaser
# https://github.com/goreleaser/goreleaser/issues/5487
#config: ./supplemental/debian/config.sh
scoops:
- ids: [beszel-agent]

View File

@@ -3,40 +3,6 @@ OS ?= $(shell go env GOOS)
ARCH ?= $(shell go env GOARCH)
# Skip building the web UI if true
SKIP_WEB ?= false
# Controls NVML/glibc agent build tag behavior:
# - auto (default): enable on linux/amd64 glibc hosts
# - true: always enable
# - false: always disable
NVML ?= auto
# Detect glibc host for local linux/amd64 builds.
HOST_GLIBC := $(shell \
if [ "$(OS)" = "linux" ] && [ "$(ARCH)" = "amd64" ]; then \
for p in /lib64/ld-linux-x86-64.so.2 /lib/x86_64-linux-gnu/ld-linux-x86-64.so.2 /lib/ld-linux-x86-64.so.2; do \
[ -e "$$p" ] && { echo true; exit 0; }; \
done; \
if command -v ldd >/dev/null 2>&1; then \
if ldd --version 2>&1 | tr '[:upper:]' '[:lower:]' | awk '/gnu libc|glibc/{found=1} END{exit !found}'; then \
echo true; \
else \
echo false; \
fi; \
else \
echo false; \
fi; \
else \
echo false; \
fi)
# Enable glibc build tag for NVML on supported Linux builds.
AGENT_GO_TAGS :=
ifeq ($(NVML),true)
AGENT_GO_TAGS := -tags glibc
else ifeq ($(NVML),auto)
ifeq ($(HOST_GLIBC),true)
AGENT_GO_TAGS := -tags glibc
endif
endif
# Set executable extension based on target OS
EXE_EXT := $(if $(filter windows,$(OS)),.exe,)
@@ -51,6 +17,7 @@ clean:
lint:
golangci-lint run
test: export GOEXPERIMENT=synctest
test:
go test -tags=testing ./...
@@ -87,7 +54,7 @@ fetch-smartctl-conditional:
# Update build-agent to include conditional .NET build
build-agent: tidy build-dotnet-conditional fetch-smartctl-conditional
GOOS=$(OS) GOARCH=$(ARCH) go build $(AGENT_GO_TAGS) -o ./build/beszel-agent_$(OS)_$(ARCH)$(EXE_EXT) -ldflags "-w -s" ./internal/cmd/agent
GOOS=$(OS) GOARCH=$(ARCH) go build -o ./build/beszel-agent_$(OS)_$(ARCH)$(EXE_EXT) -ldflags "-w -s" ./internal/cmd/agent
build-hub: tidy $(if $(filter false,$(SKIP_WEB)),build-web-ui)
GOOS=$(OS) GOARCH=$(ARCH) go build -o ./build/beszel_$(OS)_$(ARCH)$(EXE_EXT) -ldflags "-w -s" ./internal/cmd/hub
@@ -123,9 +90,9 @@ dev-hub:
dev-agent:
@if command -v entr >/dev/null 2>&1; then \
find ./internal/cmd/agent/*.go ./agent/*.go | entr -r go run $(AGENT_GO_TAGS) github.com/henrygd/beszel/internal/cmd/agent; \
find ./internal/cmd/agent/*.go ./agent/*.go | entr -r go run github.com/henrygd/beszel/internal/cmd/agent; \
else \
go run $(AGENT_GO_TAGS) github.com/henrygd/beszel/internal/cmd/agent; \
go run github.com/henrygd/beszel/internal/cmd/agent; \
fi
build-dotnet:

View File

@@ -5,17 +5,19 @@
package agent
import (
"crypto/sha256"
"encoding/hex"
"log/slog"
"os"
"path/filepath"
"strings"
"sync"
"time"
"github.com/gliderlabs/ssh"
"github.com/henrygd/beszel"
"github.com/henrygd/beszel/agent/deltatracker"
"github.com/henrygd/beszel/internal/common"
"github.com/henrygd/beszel/internal/entities/system"
"github.com/shirou/gopsutil/v4/host"
gossh "golang.org/x/crypto/ssh"
)
@@ -27,15 +29,12 @@ type Agent struct {
fsNames []string // List of filesystem device names being monitored
fsStats map[string]*system.FsStats // Keeps track of disk stats for each filesystem
diskPrev map[uint16]map[string]prevDisk // Previous disk I/O counters per cache interval
diskUsageCacheDuration time.Duration // How long to cache disk usage (to avoid waking sleeping disks)
lastDiskUsageUpdate time.Time // Last time disk usage was collected
netInterfaces map[string]struct{} // Stores all valid network interfaces
netIoStats map[uint16]system.NetIoStats // Keeps track of bandwidth usage per cache interval
netInterfaceDeltaTrackers map[uint16]*deltatracker.DeltaTracker[string, uint64] // Per-cache-time NIC delta trackers
dockerManager *dockerManager // Manages Docker API requests
sensorConfig *SensorConfig // Sensors config
systemInfo system.Info // Host system info (dynamic)
systemDetails system.Details // Host system details (static, once-per-connection)
systemInfo system.Info // Host system info
gpuManager *GPUManager // Manages GPU data
cache *systemDataCache // Cache for system stats based on cache time
connectionManager *ConnectionManager // Channel to signal connection events
@@ -61,7 +60,7 @@ func NewAgent(dataDir ...string) (agent *Agent, err error) {
agent.netIoStats = make(map[uint16]system.NetIoStats)
agent.netInterfaceDeltaTrackers = make(map[uint16]*deltatracker.DeltaTracker[string, uint64])
agent.dataDir, err = GetDataDir(dataDir...)
agent.dataDir, err = getDataDir(dataDir...)
if err != nil {
slog.Warn("Data directory not found")
} else {
@@ -70,17 +69,6 @@ func NewAgent(dataDir ...string) (agent *Agent, err error) {
agent.memCalc, _ = GetEnv("MEM_CALC")
agent.sensorConfig = agent.newSensorConfig()
// Parse disk usage cache duration (e.g., "15m", "1h") to avoid waking sleeping disks
if diskUsageCache, exists := GetEnv("DISK_USAGE_CACHE"); exists {
if duration, err := time.ParseDuration(diskUsageCache); err == nil {
agent.diskUsageCacheDuration = duration
slog.Info("DISK_USAGE_CACHE", "duration", duration)
} else {
slog.Warn("Invalid DISK_USAGE_CACHE", "err", err)
}
}
// Set up slog with a log level determined by the LOG_LEVEL env var
if logLevelStr, exists := GetEnv("LOG_LEVEL"); exists {
switch strings.ToLower(logLevelStr) {
@@ -96,21 +84,8 @@ func NewAgent(dataDir ...string) (agent *Agent, err error) {
slog.Debug(beszel.Version)
// initialize docker manager
agent.dockerManager = newDockerManager()
// initialize system info
agent.refreshSystemDetails()
// SMART_INTERVAL env var to update smart data at this interval
if smartIntervalEnv, exists := GetEnv("SMART_INTERVAL"); exists {
if duration, err := time.ParseDuration(smartIntervalEnv); err == nil && duration > 0 {
agent.systemDetails.SmartInterval = duration
slog.Info("SMART_INTERVAL", "duration", duration)
} else {
slog.Warn("Invalid SMART_INTERVAL", "err", err)
}
}
agent.initializeSystemInfo()
// initialize connection manager
agent.connectionManager = newConnectionManager(agent)
@@ -124,6 +99,9 @@ func NewAgent(dataDir ...string) (agent *Agent, err error) {
// initialize net io stats
agent.initializeNetIoStats()
// initialize docker manager
agent.dockerManager = newDockerManager(agent)
agent.systemdManager, err = newSystemdManager()
if err != nil {
slog.Debug("Systemd", "err", err)
@@ -142,7 +120,7 @@ func NewAgent(dataDir ...string) (agent *Agent, err error) {
// if debugging, print stats
if agent.debug {
slog.Debug("Stats", "data", agent.gatherStats(common.DataRequestOptions{CacheTimeMs: 60_000, IncludeDetails: true}))
slog.Debug("Stats", "data", agent.gatherStats(0))
}
return agent, nil
@@ -157,11 +135,10 @@ func GetEnv(key string) (value string, exists bool) {
return os.LookupEnv(key)
}
func (a *Agent) gatherStats(options common.DataRequestOptions) *system.CombinedData {
func (a *Agent) gatherStats(cacheTimeMs uint16) *system.CombinedData {
a.Lock()
defer a.Unlock()
cacheTimeMs := options.CacheTimeMs
data, isCached := a.cache.Get(cacheTimeMs)
if isCached {
slog.Debug("Cached data", "cacheTimeMs", cacheTimeMs)
@@ -172,12 +149,6 @@ func (a *Agent) gatherStats(options common.DataRequestOptions) *system.CombinedD
Stats: a.getSystemStats(cacheTimeMs),
Info: a.systemInfo,
}
// Include static system details only when requested
if options.IncludeDetails {
data.Details = &a.systemDetails
}
// slog.Info("System data", "data", data, "cacheTimeMs", cacheTimeMs)
if a.dockerManager != nil {
@@ -224,12 +195,37 @@ func (a *Agent) gatherStats(options common.DataRequestOptions) *system.CombinedD
return data
}
// Start initializes and starts the agent with optional WebSocket connection
// StartAgent initializes and starts the agent with optional WebSocket connection
func (a *Agent) Start(serverOptions ServerOptions) error {
a.keys = serverOptions.Keys
return a.connectionManager.Start(serverOptions)
}
func (a *Agent) getFingerprint() string {
return GetFingerprint(a.dataDir, a.systemDetails.Hostname, a.systemDetails.CpuModel)
// first look for a fingerprint in the data directory
if a.dataDir != "" {
if fp, err := os.ReadFile(filepath.Join(a.dataDir, "fingerprint")); err == nil {
return string(fp)
}
}
// if no fingerprint is found, generate one
fingerprint, err := host.HostID()
if err != nil || fingerprint == "" {
fingerprint = a.systemInfo.Hostname + a.systemInfo.CpuModel
}
// hash fingerprint
sum := sha256.Sum256([]byte(fingerprint))
fingerprint = hex.EncodeToString(sum[:24])
// save fingerprint to data directory
if a.dataDir != "" {
err = os.WriteFile(filepath.Join(a.dataDir, "fingerprint"), []byte(fingerprint), 0644)
if err != nil {
slog.Warn("Failed to save fingerprint", "err", err)
}
}
return fingerprint
}

View File

@@ -1,4 +1,5 @@
//go:build testing
// +build testing
package agent
@@ -21,7 +22,7 @@ func createTestCacheData() *system.CombinedData {
DiskTotal: 100000,
},
Info: system.Info{
AgentVersion: "0.12.0",
Hostname: "test-host",
},
Containers: []*container.Stats{
{
@@ -127,7 +128,7 @@ func TestCacheMultipleIntervals(t *testing.T) {
Mem: 16384,
},
Info: system.Info{
AgentVersion: "0.12.0",
Hostname: "test-host-2",
},
Containers: []*container.Stats{},
}
@@ -170,7 +171,7 @@ func TestCacheOverwrite(t *testing.T) {
Mem: 32768,
},
Info: system.Info{
AgentVersion: "0.12.0",
Hostname: "updated-host",
},
Containers: []*container.Stats{},
}

View File

@@ -1,4 +1,5 @@
//go:build testing
// +build testing
package agent

View File

@@ -65,7 +65,7 @@ func GetBatteryStats() (batteryPercent uint8, batteryState uint8, err error) {
continue
}
totalCapacity += bat.Full
totalCharge += min(bat.Current, bat.Full)
totalCharge += bat.Current
if bat.State.Raw >= 0 {
batteryState = uint8(bat.State.Raw)
}

View File

@@ -15,6 +15,9 @@ import (
"github.com/henrygd/beszel"
"github.com/henrygd/beszel/internal/common"
"github.com/henrygd/beszel/internal/entities/smart"
"github.com/henrygd/beszel/internal/entities/system"
"github.com/henrygd/beszel/internal/entities/systemd"
"github.com/fxamacker/cbor/v2"
"github.com/lxzan/gws"
@@ -198,7 +201,7 @@ func (client *WebSocketClient) handleAuthChallenge(msg *common.HubRequest[cbor.R
if authRequest.NeedSysInfo {
response.Name, _ = GetEnv("SYSTEM_NAME")
response.Hostname = client.agent.systemDetails.Hostname
response.Hostname = client.agent.systemInfo.Hostname
serverAddr := client.agent.connectionManager.serverOptions.Addr
_, response.Port, _ = net.SplitHostPort(serverAddr)
}
@@ -256,16 +259,40 @@ func (client *WebSocketClient) sendMessage(data any) error {
return err
}
// sendResponse sends a response with optional request ID.
// For ID-based requests, we must populate legacy typed fields for backward
// compatibility with older hubs (<= 0.17) that don't read the generic Data field.
// sendResponse sends a response with optional request ID for the new protocol
func (client *WebSocketClient) sendResponse(data any, requestID *uint32) error {
if requestID != nil {
response := newAgentResponse(data, requestID)
// New format with ID - use typed fields
response := common.AgentResponse{
Id: requestID,
}
// Set the appropriate typed field based on data type
switch v := data.(type) {
case *system.CombinedData:
response.SystemData = v
case *common.FingerprintResponse:
response.Fingerprint = v
case string:
response.String = &v
case map[string]smart.SmartData:
response.SmartData = v
case systemd.ServiceDetails:
response.ServiceInfo = v
// case []byte:
// response.RawBytes = v
// case string:
// response.RawBytes = []byte(v)
default:
// For any other type, convert to error
response.Error = fmt.Sprintf("unsupported response type: %T", data)
}
return client.sendMessage(response)
} else {
// Legacy format - send data directly
return client.sendMessage(data)
}
// Legacy format - send data directly
return client.sendMessage(data)
}
// getUserAgent returns one of two User-Agent strings based on current time.

View File

@@ -1,4 +1,5 @@
//go:build testing
// +build testing
package agent

View File

@@ -1,9 +1,9 @@
package agent
import (
"context"
"errors"
"log/slog"
"os"
"os/signal"
"syscall"
"time"
@@ -91,8 +91,8 @@ func (c *ConnectionManager) Start(serverOptions ServerOptions) error {
c.eventChan = make(chan ConnectionEvent, 1)
// signal handling for shutdown
sigCtx, stopSignals := signal.NotifyContext(context.Background(), syscall.SIGINT, syscall.SIGTERM)
defer stopSignals()
sigChan := make(chan os.Signal, 1)
signal.Notify(sigChan, syscall.SIGINT, syscall.SIGTERM)
c.startWsTicker()
c.connect()
@@ -109,8 +109,8 @@ func (c *ConnectionManager) Start(serverOptions ServerOptions) error {
_ = c.startWebSocketConnection()
case <-healthTicker:
_ = health.Update()
case <-sigCtx.Done():
slog.Info("Shutting down", "cause", context.Cause(sigCtx))
case <-sigChan:
slog.Info("Shutting down")
_ = c.agent.StopServer()
c.closeWebSocket()
return health.CleanUp()

View File

@@ -1,4 +1,5 @@
//go:build testing
// +build testing
package agent

View File

@@ -89,7 +89,10 @@ func getPerCoreCpuUsage(cacheTimeMs uint16) (system.Uint8Slice, error) {
lastTimes := lastPerCoreCpuTimes[cacheTimeMs]
// Limit to the number of cores available in both samples
length := min(len(lastTimes), len(perCoreTimes))
length := len(perCoreTimes)
if len(lastTimes) < length {
length = len(lastTimes)
}
usage := make([]uint8, length)
for i := 0; i < length; i++ {

View File

@@ -8,10 +8,10 @@ import (
"runtime"
)
// GetDataDir returns the path to the data directory for the agent and an error
// getDataDir returns the path to the data directory for the agent and an error
// if the directory is not valid. Attempts to find the optimal data directory if
// no data directories are provided.
func GetDataDir(dataDirs ...string) (string, error) {
func getDataDir(dataDirs ...string) (string, error) {
if len(dataDirs) > 0 {
return testDataDirs(dataDirs)
}

View File

@@ -1,4 +1,5 @@
//go:build testing
// +build testing
package agent
@@ -16,7 +17,7 @@ func TestGetDataDir(t *testing.T) {
// Test with explicit dataDir parameter
t.Run("explicit data dir", func(t *testing.T) {
tempDir := t.TempDir()
result, err := GetDataDir(tempDir)
result, err := getDataDir(tempDir)
require.NoError(t, err)
assert.Equal(t, tempDir, result)
})
@@ -25,7 +26,7 @@ func TestGetDataDir(t *testing.T) {
t.Run("explicit data dir - create new", func(t *testing.T) {
tempDir := t.TempDir()
newDir := filepath.Join(tempDir, "new-data-dir")
result, err := GetDataDir(newDir)
result, err := getDataDir(newDir)
require.NoError(t, err)
assert.Equal(t, newDir, result)
@@ -51,7 +52,7 @@ func TestGetDataDir(t *testing.T) {
os.Setenv("BESZEL_AGENT_DATA_DIR", tempDir)
result, err := GetDataDir()
result, err := getDataDir()
require.NoError(t, err)
assert.Equal(t, tempDir, result)
})
@@ -59,7 +60,7 @@ func TestGetDataDir(t *testing.T) {
// Test with invalid explicit dataDir
t.Run("invalid explicit data dir", func(t *testing.T) {
invalidPath := "/invalid/path/that/cannot/be/created"
_, err := GetDataDir(invalidPath)
_, err := getDataDir(invalidPath)
assert.Error(t, err)
})
@@ -78,7 +79,7 @@ func TestGetDataDir(t *testing.T) {
// This will try platform-specific defaults, which may or may not work
// We're mainly testing that it doesn't panic and returns some result
result, err := GetDataDir()
result, err := getDataDir()
// We don't assert success/failure here since it depends on system permissions
// Just verify we get a string result if no error
if err == nil {

View File

@@ -26,15 +26,6 @@ func parseFilesystemEntry(entry string) (device, customName string) {
return device, customName
}
func isDockerSpecialMountpoint(mountpoint string) bool {
switch mountpoint {
case "/etc/hosts", "/etc/resolv.conf", "/etc/hostname":
return true
default:
return false
}
}
// Sets up the filesystems to monitor for disk usage and I/O.
func (a *Agent) initializeDiskInfo() {
filesystem, _ := GetEnv("FILESYSTEM")
@@ -78,15 +69,11 @@ func (a *Agent) initializeDiskInfo() {
if _, exists := a.fsStats[key]; !exists {
if root {
slog.Info("Detected root device", "name", key)
// Check if root device is in /proc/diskstats. Do not guess a
// fallback device for root: that can misattribute root I/O to a
// different disk while usage remains tied to root mountpoint.
// Check if root device is in /proc/diskstats, use fallback if not
if _, ioMatch = diskIoCounters[key]; !ioMatch {
if matchedKey, match := findIoDevice(filesystem, diskIoCounters); match {
key = matchedKey
ioMatch = true
} else {
slog.Warn("Root I/O unmapped; set FILESYSTEM", "device", device, "mountpoint", mountpoint)
key, ioMatch = findIoDevice(filesystem, diskIoCounters, a.fsStats)
if !ioMatch {
slog.Info("Using I/O fallback", "device", device, "mountpoint", mountpoint, "fallback", key)
}
}
} else {
@@ -127,7 +114,7 @@ func (a *Agent) initializeDiskInfo() {
// Add EXTRA_FILESYSTEMS env var values to fsStats
if extraFilesystems, exists := GetEnv("EXTRA_FILESYSTEMS"); exists {
for fsEntry := range strings.SplitSeq(extraFilesystems, ",") {
for _, fsEntry := range strings.Split(extraFilesystems, ",") {
// Parse custom name from format: device__customname
fs, customName := parseFilesystemEntry(fsEntry)
@@ -154,8 +141,8 @@ func (a *Agent) initializeDiskInfo() {
for _, p := range partitions {
// fmt.Println(p.Device, p.Mountpoint)
// Binary root fallback or docker root fallback
if !hasRoot && (p.Mountpoint == rootMountPoint || (isDockerSpecialMountpoint(p.Mountpoint) && strings.HasPrefix(p.Device, "/dev"))) {
fs, match := findIoDevice(filepath.Base(p.Device), diskIoCounters)
if !hasRoot && (p.Mountpoint == rootMountPoint || (p.Mountpoint == "/etc/hosts" && strings.HasPrefix(p.Device, "/dev"))) {
fs, match := findIoDevice(filepath.Base(p.Device), diskIoCounters, a.fsStats)
if match {
addFsStat(fs, p.Mountpoint, true)
hasRoot = true
@@ -189,26 +176,33 @@ func (a *Agent) initializeDiskInfo() {
// If no root filesystem set, use fallback
if !hasRoot {
rootKey := filepath.Base(rootMountPoint)
if _, exists := a.fsStats[rootKey]; exists {
rootKey = "root"
}
slog.Warn("Root device not detected; root I/O disabled", "mountpoint", rootMountPoint)
a.fsStats[rootKey] = &system.FsStats{Root: true, Mountpoint: rootMountPoint}
rootDevice, _ := findIoDevice(filepath.Base(filesystem), diskIoCounters, a.fsStats)
slog.Info("Root disk", "mountpoint", rootMountPoint, "io", rootDevice)
a.fsStats[rootDevice] = &system.FsStats{Root: true, Mountpoint: rootMountPoint}
}
a.initializeDiskIoStats(diskIoCounters)
}
// Returns matching device from /proc/diskstats.
// Returns matching device from /proc/diskstats,
// or the device with the most reads if no match is found.
// bool is true if a match was found.
func findIoDevice(filesystem string, diskIoCounters map[string]disk.IOCountersStat) (string, bool) {
func findIoDevice(filesystem string, diskIoCounters map[string]disk.IOCountersStat, fsStats map[string]*system.FsStats) (string, bool) {
var maxReadBytes uint64
maxReadDevice := "/"
for _, d := range diskIoCounters {
if d.Name == filesystem || (d.Label != "" && d.Label == filesystem) {
return d.Name, true
}
if d.ReadBytes > maxReadBytes {
// don't use if device already exists in fsStats
if _, exists := fsStats[d.Name]; !exists {
maxReadBytes = d.ReadBytes
maxReadDevice = d.Name
}
}
}
return "", false
return maxReadDevice, false
}
// Sets start values for disk I/O stats.
@@ -231,19 +225,8 @@ func (a *Agent) initializeDiskIoStats(diskIoCounters map[string]disk.IOCountersS
// Updates disk usage statistics for all monitored filesystems
func (a *Agent) updateDiskUsage(systemStats *system.Stats) {
// Check if we should skip extra filesystem collection to avoid waking sleeping disks.
// Root filesystem is always updated since it can't be sleeping while the agent runs.
// Always collect on first call (lastDiskUsageUpdate is zero) or if caching is disabled.
cacheExtraFs := a.diskUsageCacheDuration > 0 &&
!a.lastDiskUsageUpdate.IsZero() &&
time.Since(a.lastDiskUsageUpdate) < a.diskUsageCacheDuration
// disk usage
for _, stats := range a.fsStats {
// Skip non-root filesystems if caching is active
if cacheExtraFs && !stats.Root {
continue
}
if d, err := disk.Usage(stats.Mountpoint); err == nil {
stats.DiskTotal = bytesToGigabytes(d.Total)
stats.DiskUsed = bytesToGigabytes(d.Used)
@@ -261,11 +244,6 @@ func (a *Agent) updateDiskUsage(systemStats *system.Stats) {
stats.TotalWrite = 0
}
}
// Update the last disk usage update time when we've collected extra filesystems
if !cacheExtraFs {
a.lastDiskUsageUpdate = time.Now()
}
}
// Updates disk I/O statistics for all monitored filesystems

View File

@@ -1,4 +1,5 @@
//go:build testing
// +build testing
package agent
@@ -6,7 +7,6 @@ import (
"os"
"strings"
"testing"
"time"
"github.com/henrygd/beszel/internal/entities/system"
"github.com/shirou/gopsutil/v4/disk"
@@ -93,62 +93,6 @@ func TestParseFilesystemEntry(t *testing.T) {
}
}
func TestFindIoDevice(t *testing.T) {
t.Run("matches by device name", func(t *testing.T) {
ioCounters := map[string]disk.IOCountersStat{
"sda": {Name: "sda"},
"sdb": {Name: "sdb"},
}
device, ok := findIoDevice("sdb", ioCounters)
assert.True(t, ok)
assert.Equal(t, "sdb", device)
})
t.Run("matches by device label", func(t *testing.T) {
ioCounters := map[string]disk.IOCountersStat{
"sda": {Name: "sda", Label: "rootfs"},
"sdb": {Name: "sdb"},
}
device, ok := findIoDevice("rootfs", ioCounters)
assert.True(t, ok)
assert.Equal(t, "sda", device)
})
t.Run("returns no fallback when not found", func(t *testing.T) {
ioCounters := map[string]disk.IOCountersStat{
"sda": {Name: "sda"},
"sdb": {Name: "sdb"},
}
device, ok := findIoDevice("nvme0n1p1", ioCounters)
assert.False(t, ok)
assert.Equal(t, "", device)
})
}
func TestIsDockerSpecialMountpoint(t *testing.T) {
testCases := []struct {
name string
mountpoint string
expected bool
}{
{name: "hosts", mountpoint: "/etc/hosts", expected: true},
{name: "resolv", mountpoint: "/etc/resolv.conf", expected: true},
{name: "hostname", mountpoint: "/etc/hostname", expected: true},
{name: "root", mountpoint: "/", expected: false},
{name: "passwd", mountpoint: "/etc/passwd", expected: false},
{name: "extra-filesystem", mountpoint: "/extra-filesystems/sda1", expected: false},
}
for _, tc := range testCases {
t.Run(tc.name, func(t *testing.T) {
assert.Equal(t, tc.expected, isDockerSpecialMountpoint(tc.mountpoint))
})
}
}
func TestInitializeDiskInfoWithCustomNames(t *testing.T) {
// Set up environment variables
oldEnv := os.Getenv("EXTRA_FILESYSTEMS")
@@ -289,86 +233,3 @@ func TestExtraFsKeyGeneration(t *testing.T) {
})
}
}
func TestDiskUsageCaching(t *testing.T) {
t.Run("caching disabled updates all filesystems", func(t *testing.T) {
agent := &Agent{
fsStats: map[string]*system.FsStats{
"sda": {Root: true, Mountpoint: "/"},
"sdb": {Root: false, Mountpoint: "/mnt/storage"},
},
diskUsageCacheDuration: 0, // caching disabled
}
var stats system.Stats
agent.updateDiskUsage(&stats)
// Both should be updated (non-zero values from disk.Usage)
// Root stats should be populated in systemStats
assert.True(t, agent.lastDiskUsageUpdate.IsZero() || !agent.lastDiskUsageUpdate.IsZero(),
"lastDiskUsageUpdate should be set when caching is disabled")
})
t.Run("caching enabled always updates root filesystem", func(t *testing.T) {
agent := &Agent{
fsStats: map[string]*system.FsStats{
"sda": {Root: true, Mountpoint: "/", DiskTotal: 100, DiskUsed: 50},
"sdb": {Root: false, Mountpoint: "/mnt/storage", DiskTotal: 200, DiskUsed: 100},
},
diskUsageCacheDuration: 1 * time.Hour,
lastDiskUsageUpdate: time.Now(), // cache is fresh
}
// Store original extra fs values
originalExtraTotal := agent.fsStats["sdb"].DiskTotal
originalExtraUsed := agent.fsStats["sdb"].DiskUsed
var stats system.Stats
agent.updateDiskUsage(&stats)
// Root should be updated (systemStats populated from disk.Usage call)
// We can't easily check if disk.Usage was called, but we verify the flow works
// Extra filesystem should retain cached values (not reset)
assert.Equal(t, originalExtraTotal, agent.fsStats["sdb"].DiskTotal,
"extra filesystem DiskTotal should be unchanged when cached")
assert.Equal(t, originalExtraUsed, agent.fsStats["sdb"].DiskUsed,
"extra filesystem DiskUsed should be unchanged when cached")
})
t.Run("first call always updates all filesystems", func(t *testing.T) {
agent := &Agent{
fsStats: map[string]*system.FsStats{
"sda": {Root: true, Mountpoint: "/"},
"sdb": {Root: false, Mountpoint: "/mnt/storage"},
},
diskUsageCacheDuration: 1 * time.Hour,
// lastDiskUsageUpdate is zero (first call)
}
var stats system.Stats
agent.updateDiskUsage(&stats)
// After first call, lastDiskUsageUpdate should be set
assert.False(t, agent.lastDiskUsageUpdate.IsZero(),
"lastDiskUsageUpdate should be set after first call")
})
t.Run("expired cache updates extra filesystems", func(t *testing.T) {
agent := &Agent{
fsStats: map[string]*system.FsStats{
"sda": {Root: true, Mountpoint: "/"},
"sdb": {Root: false, Mountpoint: "/mnt/storage"},
},
diskUsageCacheDuration: 1 * time.Millisecond,
lastDiskUsageUpdate: time.Now().Add(-1 * time.Second), // cache expired
}
var stats system.Stats
agent.updateDiskUsage(&stats)
// lastDiskUsageUpdate should be refreshed since cache expired
assert.True(t, time.Since(agent.lastDiskUsageUpdate) < time.Second,
"lastDiskUsageUpdate should be refreshed when cache expires")
})
}

View File

@@ -1,7 +1,6 @@
package agent
import (
"bufio"
"bytes"
"context"
"encoding/binary"
@@ -15,7 +14,6 @@ import (
"net/url"
"os"
"path"
"regexp"
"strings"
"sync"
"time"
@@ -26,11 +24,6 @@ import (
"github.com/blang/semver"
)
// ansiEscapePattern matches ANSI escape sequences (colors, cursor movement, etc.)
// This includes CSI sequences like \x1b[...m and simple escapes like \x1b[K
var ansiEscapePattern = regexp.MustCompile(`\x1b\[[0-9;]*[a-zA-Z]|\x1b\][^\x07]*\x07|\x1b[@-Z\\-_]`)
var dockerContainerIDPattern = regexp.MustCompile(`^[a-fA-F0-9]{12,64}$`)
const (
// Docker API timeout in milliseconds
dockerTimeoutMs = 2100
@@ -62,7 +55,6 @@ type dockerManager struct {
decoder *json.Decoder // Reusable JSON decoder that reads from buf
apiStats *container.ApiStats // Reusable API stats object
excludeContainers []string // Patterns to exclude containers by name
usingPodman bool // Whether the Docker Engine API is running on Podman
// Cache-time-aware tracking for CPU stats (similar to cpu.go)
// Maps cache time intervals to container-specific CPU usage tracking
@@ -74,7 +66,6 @@ type dockerManager struct {
// cacheTimeMs -> DeltaTracker for network bytes sent/received
networkSentTrackers map[uint16]*deltatracker.DeltaTracker[string, uint64]
networkRecvTrackers map[uint16]*deltatracker.DeltaTracker[string, uint64]
retrySleep func(time.Duration)
}
// userAgentRoundTripper is a custom http.RoundTripper that adds a User-Agent header to all requests
@@ -338,8 +329,6 @@ func validateCpuPercentage(cpuPct float64, containerName string) error {
func updateContainerStatsValues(stats *container.Stats, cpuPct float64, usedMemory uint64, sent_delta, recv_delta uint64, readTime time.Time) {
stats.Cpu = twoDecimals(cpuPct)
stats.Mem = bytesToMegabytes(float64(usedMemory))
stats.Bandwidth = [2]uint64{sent_delta, recv_delta}
// TODO(0.19+): stop populating NetworkSent/NetworkRecv (deprecated in 0.18.3)
stats.NetworkSent = bytesToMegabytes(float64(sent_delta))
stats.NetworkRecv = bytesToMegabytes(float64(recv_delta))
stats.PrevReadTime = readTime
@@ -408,8 +397,6 @@ func (dm *dockerManager) updateContainerStats(ctr *container.ApiInfo, cacheTimeM
// reset current stats
stats.Cpu = 0
stats.Mem = 0
stats.Bandwidth = [2]uint64{0, 0}
// TODO(0.19+): stop populating NetworkSent/NetworkRecv (deprecated in 0.18.3)
stats.NetworkSent = 0
stats.NetworkRecv = 0
@@ -486,7 +473,7 @@ func (dm *dockerManager) deleteContainerStatsSync(id string) {
}
// Creates a new http client for Docker or Podman API
func newDockerManager() *dockerManager {
func newDockerManager(a *Agent) *dockerManager {
dockerHost, exists := GetEnv("DOCKER_HOST")
if exists {
// return nil if set to empty string
@@ -568,17 +555,16 @@ func newDockerManager() *dockerManager {
lastCpuReadTime: make(map[uint16]map[string]time.Time),
networkSentTrackers: make(map[uint16]*deltatracker.DeltaTracker[string, uint64]),
networkRecvTrackers: make(map[uint16]*deltatracker.DeltaTracker[string, uint64]),
retrySleep: time.Sleep,
}
// If using podman, return client
if strings.Contains(dockerHost, "podman") {
manager.usingPodman = true
a.systemInfo.Podman = true
manager.goodDockerVersion = true
return manager
}
// run version check in goroutine to avoid blocking (server may not be ready and requires retries)
// this can take up to 5 seconds with retry, so run in goroutine
go manager.checkDockerVersion()
// give version check a chance to complete before returning
@@ -598,18 +584,18 @@ func (dm *dockerManager) checkDockerVersion() {
const versionMaxTries = 2
for i := 1; i <= versionMaxTries; i++ {
resp, err = dm.client.Get("http://localhost/version")
if err == nil && resp.StatusCode == http.StatusOK {
if err == nil {
break
}
if resp != nil {
resp.Body.Close()
}
if i < versionMaxTries {
slog.Debug("Failed to get Docker version; retrying", "attempt", i, "err", err, "response", resp)
dm.retrySleep(5 * time.Second)
slog.Debug("Failed to get Docker version; retrying", "attempt", i, "error", err)
time.Sleep(5 * time.Second)
}
}
if err != nil || resp.StatusCode != http.StatusOK {
if err != nil {
return
}
if err := dm.decode(resp, &versionInfo); err != nil {
@@ -651,34 +637,9 @@ func getDockerHost() string {
return scheme + socks[0]
}
func validateContainerID(containerID string) error {
if !dockerContainerIDPattern.MatchString(containerID) {
return fmt.Errorf("invalid container id")
}
return nil
}
func buildDockerContainerEndpoint(containerID, action string, query url.Values) (string, error) {
if err := validateContainerID(containerID); err != nil {
return "", err
}
u := &url.URL{
Scheme: "http",
Host: "localhost",
Path: fmt.Sprintf("/containers/%s/%s", url.PathEscape(containerID), action),
}
if len(query) > 0 {
u.RawQuery = query.Encode()
}
return u.String(), nil
}
// getContainerInfo fetches the inspection data for a container
func (dm *dockerManager) getContainerInfo(ctx context.Context, containerID string) ([]byte, error) {
endpoint, err := buildDockerContainerEndpoint(containerID, "json", nil)
if err != nil {
return nil, err
}
endpoint := fmt.Sprintf("http://localhost/containers/%s/json", containerID)
req, err := http.NewRequestWithContext(ctx, http.MethodGet, endpoint, nil)
if err != nil {
return nil, err
@@ -709,15 +670,7 @@ func (dm *dockerManager) getContainerInfo(ctx context.Context, containerID strin
// getLogs fetches the logs for a container
func (dm *dockerManager) getLogs(ctx context.Context, containerID string) (string, error) {
query := url.Values{
"stdout": []string{"1"},
"stderr": []string{"1"},
"tail": []string{fmt.Sprintf("%d", dockerLogsTail)},
}
endpoint, err := buildDockerContainerEndpoint(containerID, "logs", query)
if err != nil {
return "", err
}
endpoint := fmt.Sprintf("http://localhost/containers/%s/logs?stdout=1&stderr=1&tail=%d", containerID, dockerLogsTail)
req, err := http.NewRequestWithContext(ctx, http.MethodGet, endpoint, nil)
if err != nil {
return "", err
@@ -735,52 +688,17 @@ func (dm *dockerManager) getLogs(ctx context.Context, containerID string) (strin
}
var builder strings.Builder
contentType := resp.Header.Get("Content-Type")
multiplexed := strings.HasSuffix(contentType, "multiplexed-stream")
logReader := io.Reader(resp.Body)
if !multiplexed {
// Podman may return multiplexed logs without Content-Type. Sniff the first frame header
// with a small buffered reader only when the header check fails.
bufferedReader := bufio.NewReaderSize(resp.Body, 8)
multiplexed = detectDockerMultiplexedStream(bufferedReader)
logReader = bufferedReader
}
if err := decodeDockerLogStream(logReader, &builder, multiplexed); err != nil {
if err := decodeDockerLogStream(resp.Body, &builder); err != nil {
return "", err
}
// Strip ANSI escape sequences from logs for clean display in web UI
logs := builder.String()
if strings.Contains(logs, "\x1b") {
logs = ansiEscapePattern.ReplaceAllString(logs, "")
}
return logs, nil
return builder.String(), nil
}
func detectDockerMultiplexedStream(reader *bufio.Reader) bool {
const headerSize = 8
header, err := reader.Peek(headerSize)
if err != nil {
return false
}
if header[0] != 0x01 && header[0] != 0x02 {
return false
}
// Docker's stream framing header reserves bytes 1-3 as zero.
if header[1] != 0 || header[2] != 0 || header[3] != 0 {
return false
}
frameLen := binary.BigEndian.Uint32(header[4:])
return frameLen <= maxLogFrameSize
}
func decodeDockerLogStream(reader io.Reader, builder *strings.Builder, multiplexed bool) error {
if !multiplexed {
_, err := io.Copy(builder, io.LimitReader(reader, maxTotalLogSize))
return err
}
func decodeDockerLogStream(reader io.Reader, builder *strings.Builder) error {
const headerSize = 8
var header [headerSize]byte
buf := make([]byte, 0, dockerLogsTail*200)
totalBytesRead := 0
for {
@@ -804,37 +722,36 @@ func decodeDockerLogStream(reader io.Reader, builder *strings.Builder, multiplex
// Check if reading this frame would exceed total log size limit
if totalBytesRead+int(frameLen) > maxTotalLogSize {
// Read and discard remaining data to avoid blocking
_, _ = io.CopyN(io.Discard, reader, int64(frameLen))
_, _ = io.Copy(io.Discard, io.LimitReader(reader, int64(frameLen)))
slog.Debug("Truncating logs: limit reached", "read", totalBytesRead, "limit", maxTotalLogSize)
return nil
}
n, err := io.CopyN(builder, reader, int64(frameLen))
if err != nil {
buf = allocateBuffer(buf, int(frameLen))
if _, err := io.ReadFull(reader, buf[:frameLen]); err != nil {
if errors.Is(err, io.EOF) || errors.Is(err, io.ErrUnexpectedEOF) {
if len(buf) > 0 {
builder.Write(buf[:min(int(frameLen), len(buf))])
}
return nil
}
return err
}
totalBytesRead += int(n)
builder.Write(buf[:frameLen])
totalBytesRead += int(frameLen)
}
}
// GetHostInfo fetches the system info from Docker
func (dm *dockerManager) GetHostInfo() (info container.HostInfo, err error) {
resp, err := dm.client.Get("http://localhost/info")
if err != nil {
return info, err
func allocateBuffer(current []byte, needed int) []byte {
if cap(current) >= needed {
return current[:needed]
}
defer resp.Body.Close()
if err := json.NewDecoder(resp.Body).Decode(&info); err != nil {
return info, err
}
return info, nil
return make([]byte, needed)
}
func (dm *dockerManager) IsPodman() bool {
return dm.usingPodman
func min(a, b int) int {
if a < b {
return a
}
return b
}

View File

@@ -1,17 +1,11 @@
//go:build testing
// +build testing
package agent
import (
"bytes"
"context"
"encoding/json"
"errors"
"fmt"
"io"
"net"
"net/http"
"net/http/httptest"
"os"
"strings"
"testing"
@@ -25,37 +19,6 @@ import (
var defaultCacheTimeMs = uint16(60_000)
type recordingRoundTripper struct {
statusCode int
body string
contentType string
called bool
lastPath string
lastQuery map[string]string
}
func (rt *recordingRoundTripper) RoundTrip(req *http.Request) (*http.Response, error) {
rt.called = true
rt.lastPath = req.URL.EscapedPath()
rt.lastQuery = map[string]string{}
for key, values := range req.URL.Query() {
if len(values) > 0 {
rt.lastQuery[key] = values[0]
}
}
resp := &http.Response{
StatusCode: rt.statusCode,
Status: "200 OK",
Header: make(http.Header),
Body: io.NopCloser(strings.NewReader(rt.body)),
Request: req,
}
if rt.contentType != "" {
resp.Header.Set("Content-Type", rt.contentType)
}
return resp, nil
}
// cycleCpuDeltas cycles the CPU tracking data for a specific cache time interval
func (dm *dockerManager) cycleCpuDeltas(cacheTimeMs uint16) {
// Clear the CPU tracking maps for this cache time interval
@@ -147,72 +110,6 @@ func TestCalculateMemoryUsage(t *testing.T) {
}
}
func TestBuildDockerContainerEndpoint(t *testing.T) {
t.Run("valid container ID builds escaped endpoint", func(t *testing.T) {
endpoint, err := buildDockerContainerEndpoint("0123456789ab", "json", nil)
require.NoError(t, err)
assert.Equal(t, "http://localhost/containers/0123456789ab/json", endpoint)
})
t.Run("invalid container ID is rejected", func(t *testing.T) {
_, err := buildDockerContainerEndpoint("../../version", "json", nil)
require.Error(t, err)
assert.Contains(t, err.Error(), "invalid container id")
})
}
func TestContainerDetailsRequestsValidateContainerID(t *testing.T) {
rt := &recordingRoundTripper{
statusCode: 200,
body: `{"Config":{"Env":["SECRET=1"]}}`,
}
dm := &dockerManager{
client: &http.Client{Transport: rt},
}
_, err := dm.getContainerInfo(context.Background(), "../version")
require.Error(t, err)
assert.Contains(t, err.Error(), "invalid container id")
assert.False(t, rt.called, "request should be rejected before dispatching to Docker API")
}
func TestContainerDetailsRequestsUseExpectedDockerPaths(t *testing.T) {
t.Run("container info uses container json endpoint", func(t *testing.T) {
rt := &recordingRoundTripper{
statusCode: 200,
body: `{"Config":{"Env":["SECRET=1"]},"Name":"demo"}`,
}
dm := &dockerManager{
client: &http.Client{Transport: rt},
}
body, err := dm.getContainerInfo(context.Background(), "0123456789ab")
require.NoError(t, err)
assert.True(t, rt.called)
assert.Equal(t, "/containers/0123456789ab/json", rt.lastPath)
assert.NotContains(t, string(body), "SECRET=1", "sensitive env vars should be removed")
})
t.Run("container logs uses expected endpoint and query params", func(t *testing.T) {
rt := &recordingRoundTripper{
statusCode: 200,
body: "line1\nline2\n",
}
dm := &dockerManager{
client: &http.Client{Transport: rt},
}
logs, err := dm.getLogs(context.Background(), "abcdef123456")
require.NoError(t, err)
assert.True(t, rt.called)
assert.Equal(t, "/containers/abcdef123456/logs", rt.lastPath)
assert.Equal(t, "1", rt.lastQuery["stdout"])
assert.Equal(t, "1", rt.lastQuery["stderr"])
assert.Equal(t, "200", rt.lastQuery["tail"])
assert.Equal(t, "line1\nline2\n", logs)
})
}
func TestValidateCpuPercentage(t *testing.T) {
tests := []struct {
name string
@@ -287,12 +184,11 @@ func TestUpdateContainerStatsValues(t *testing.T) {
// Check memory (should be converted to MB: 1048576 bytes = 1 MB)
assert.Equal(t, 1.0, stats.Mem)
// Check bandwidth (raw bytes)
assert.Equal(t, [2]uint64{524288, 262144}, stats.Bandwidth)
// Check network sent (should be converted to MB: 524288 bytes = 0.5 MB)
assert.Equal(t, 0.5, stats.NetworkSent)
// Deprecated fields still populated for backward compatibility with older hubs
assert.Equal(t, 0.5, stats.NetworkSent) // 524288 bytes = 0.5 MB
assert.Equal(t, 0.25, stats.NetworkRecv) // 262144 bytes = 0.25 MB
// Check network recv (should be converted to MB: 262144 bytes = 0.25 MB)
assert.Equal(t, 0.25, stats.NetworkRecv)
// Check read time
assert.Equal(t, testTime, stats.PrevReadTime)
@@ -482,117 +378,6 @@ func TestDockerManagerCreation(t *testing.T) {
assert.NotNil(t, dm.networkRecvTrackers)
}
func TestCheckDockerVersion(t *testing.T) {
tests := []struct {
name string
responses []struct {
statusCode int
body string
}
expectedGood bool
expectedRequests int
}{
{
name: "200 with good version on first try",
responses: []struct {
statusCode int
body string
}{
{http.StatusOK, `{"Version":"25.0.1"}`},
},
expectedGood: true,
expectedRequests: 1,
},
{
name: "200 with old version on first try",
responses: []struct {
statusCode int
body string
}{
{http.StatusOK, `{"Version":"24.0.7"}`},
},
expectedGood: false,
expectedRequests: 1,
},
{
name: "non-200 then 200 with good version",
responses: []struct {
statusCode int
body string
}{
{http.StatusServiceUnavailable, `"not ready"`},
{http.StatusOK, `{"Version":"25.1.0"}`},
},
expectedGood: true,
expectedRequests: 2,
},
{
name: "non-200 on all retries",
responses: []struct {
statusCode int
body string
}{
{http.StatusInternalServerError, `"error"`},
{http.StatusUnauthorized, `"error"`},
},
expectedGood: false,
expectedRequests: 2,
},
}
for _, tt := range tests {
t.Run(tt.name, func(t *testing.T) {
requestCount := 0
server := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
idx := requestCount
requestCount++
if idx >= len(tt.responses) {
idx = len(tt.responses) - 1
}
w.WriteHeader(tt.responses[idx].statusCode)
fmt.Fprint(w, tt.responses[idx].body)
}))
defer server.Close()
dm := &dockerManager{
client: &http.Client{
Transport: &http.Transport{
DialContext: func(_ context.Context, network, _ string) (net.Conn, error) {
return net.Dial(network, server.Listener.Addr().String())
},
},
},
retrySleep: func(time.Duration) {},
}
dm.checkDockerVersion()
assert.Equal(t, tt.expectedGood, dm.goodDockerVersion)
assert.Equal(t, tt.expectedRequests, requestCount)
})
}
t.Run("request error on all retries", func(t *testing.T) {
requestCount := 0
dm := &dockerManager{
client: &http.Client{
Transport: &http.Transport{
DialContext: func(_ context.Context, _, _ string) (net.Conn, error) {
requestCount++
return nil, errors.New("connection refused")
},
},
},
retrySleep: func(time.Duration) {},
}
dm.checkDockerVersion()
assert.False(t, dm.goodDockerVersion)
assert.Equal(t, 2, requestCount)
})
}
func TestCycleCpuDeltas(t *testing.T) {
dm := &dockerManager{
lastCpuContainer: map[uint16]map[string]uint64{
@@ -742,10 +527,8 @@ func TestContainerStatsInitialization(t *testing.T) {
assert.Equal(t, 45.67, stats.Cpu)
assert.Equal(t, 2.0, stats.Mem)
assert.Equal(t, [2]uint64{1048576, 524288}, stats.Bandwidth)
// Deprecated fields still populated for backward compatibility with older hubs
assert.Equal(t, 1.0, stats.NetworkSent) // 1048576 bytes = 1 MB
assert.Equal(t, 0.5, stats.NetworkRecv) // 524288 bytes = 0.5 MB
assert.Equal(t, 1.0, stats.NetworkSent)
assert.Equal(t, 0.5, stats.NetworkRecv)
assert.Equal(t, testTime, stats.PrevReadTime)
}
@@ -906,49 +689,11 @@ func TestContainerStatsEndToEndWithRealData(t *testing.T) {
assert.Equal(t, cpuPct, testStats.Cpu)
assert.Equal(t, bytesToMegabytes(float64(usedMemory)), testStats.Mem)
assert.Equal(t, [2]uint64{1000000, 500000}, testStats.Bandwidth)
// Deprecated fields still populated for backward compatibility with older hubs
assert.Equal(t, bytesToMegabytes(1000000), testStats.NetworkSent)
assert.Equal(t, bytesToMegabytes(500000), testStats.NetworkRecv)
assert.Equal(t, testTime, testStats.PrevReadTime)
}
func TestGetLogsDetectsMultiplexedWithoutContentType(t *testing.T) {
// Docker multiplexed frame: [stream][0,0,0][len(4 bytes BE)][payload]
frame := []byte{
0x01, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x05,
'H', 'e', 'l', 'l', 'o',
}
rt := &recordingRoundTripper{
statusCode: 200,
body: string(frame),
// Intentionally omit content type to simulate Podman behavior.
}
dm := &dockerManager{
client: &http.Client{Transport: rt},
}
logs, err := dm.getLogs(context.Background(), "abcdef123456")
require.NoError(t, err)
assert.Equal(t, "Hello", logs)
}
func TestGetLogsDoesNotMisclassifyRawStreamAsMultiplexed(t *testing.T) {
// Starts with 0x01, but doesn't match Docker frame signature (reserved bytes aren't all zero).
raw := []byte{0x01, 0x02, 0x03, 0x04, 'r', 'a', 'w'}
rt := &recordingRoundTripper{
statusCode: 200,
body: string(raw),
}
dm := &dockerManager{
client: &http.Client{Transport: rt},
}
logs, err := dm.getLogs(context.Background(), "abcdef123456")
require.NoError(t, err)
assert.Equal(t, raw, []byte(logs))
}
func TestEdgeCasesWithRealData(t *testing.T) {
// Test with minimal container stats
minimalStats := &container.ApiStats{
@@ -1057,24 +802,6 @@ func TestNetworkRateCalculationFormula(t *testing.T) {
}
}
func TestGetHostInfo(t *testing.T) {
data, err := os.ReadFile("test-data/system_info.json")
require.NoError(t, err)
var info container.HostInfo
err = json.Unmarshal(data, &info)
require.NoError(t, err)
assert.Equal(t, "6.8.0-31-generic", info.KernelVersion)
assert.Equal(t, "Ubuntu 24.04 LTS", info.OperatingSystem)
// assert.Equal(t, "24.04", info.OSVersion)
// assert.Equal(t, "linux", info.OSType)
// assert.Equal(t, "x86_64", info.Architecture)
assert.EqualValues(t, 4, info.NCPU)
assert.EqualValues(t, 2095882240, info.MemTotal)
// assert.Equal(t, "27.0.1", info.ServerVersion)
}
func TestDeltaTrackerCacheTimeIsolation(t *testing.T) {
// Test that different cache times have separate DeltaTracker instances
dm := &dockerManager{
@@ -1205,7 +932,6 @@ func TestDecodeDockerLogStream(t *testing.T) {
input []byte
expected string
expectError bool
multiplexed bool
}{
{
name: "simple log entry",
@@ -1216,7 +942,6 @@ func TestDecodeDockerLogStream(t *testing.T) {
},
expected: "Hello World",
expectError: false,
multiplexed: true,
},
{
name: "multiple frames",
@@ -1230,7 +955,6 @@ func TestDecodeDockerLogStream(t *testing.T) {
},
expected: "HelloWorld",
expectError: false,
multiplexed: true,
},
{
name: "zero length frame",
@@ -1243,20 +967,12 @@ func TestDecodeDockerLogStream(t *testing.T) {
},
expected: "Hello",
expectError: false,
multiplexed: true,
},
{
name: "empty input",
input: []byte{},
expected: "",
expectError: false,
multiplexed: true,
},
{
name: "raw stream (not multiplexed)",
input: []byte("raw log content"),
expected: "raw log content",
multiplexed: false,
},
}
@@ -1264,7 +980,7 @@ func TestDecodeDockerLogStream(t *testing.T) {
t.Run(tt.name, func(t *testing.T) {
reader := bytes.NewReader(tt.input)
var builder strings.Builder
err := decodeDockerLogStream(reader, &builder, tt.multiplexed)
err := decodeDockerLogStream(reader, &builder)
if tt.expectError {
assert.Error(t, err)
@@ -1288,7 +1004,7 @@ func TestDecodeDockerLogStreamMemoryProtection(t *testing.T) {
reader := bytes.NewReader(input)
var builder strings.Builder
err := decodeDockerLogStream(reader, &builder, true)
err := decodeDockerLogStream(reader, &builder)
assert.Error(t, err)
assert.Contains(t, err.Error(), "log frame size")
@@ -1322,7 +1038,7 @@ func TestDecodeDockerLogStreamMemoryProtection(t *testing.T) {
reader := bytes.NewReader(input)
var builder strings.Builder
err := decodeDockerLogStream(reader, &builder, true)
err := decodeDockerLogStream(reader, &builder)
// Should complete without error (graceful truncation)
assert.NoError(t, err)
@@ -1337,6 +1053,53 @@ func TestDecodeDockerLogStreamMemoryProtection(t *testing.T) {
})
}
func TestAllocateBuffer(t *testing.T) {
tests := []struct {
name string
currentCap int
needed int
expectedCap int
shouldRealloc bool
}{
{
name: "buffer has enough capacity",
currentCap: 1024,
needed: 512,
expectedCap: 1024,
shouldRealloc: false,
},
{
name: "buffer needs reallocation",
currentCap: 512,
needed: 1024,
expectedCap: 1024,
shouldRealloc: true,
},
{
name: "buffer needs exact size",
currentCap: 1024,
needed: 1024,
expectedCap: 1024,
shouldRealloc: false,
},
}
for _, tt := range tests {
t.Run(tt.name, func(t *testing.T) {
current := make([]byte, 0, tt.currentCap)
result := allocateBuffer(current, tt.needed)
assert.Equal(t, tt.needed, len(result))
assert.GreaterOrEqual(t, cap(result), tt.expectedCap)
if tt.shouldRealloc {
// If reallocation was needed, capacity should be at least the needed size
assert.GreaterOrEqual(t, cap(result), tt.needed)
}
})
}
}
func TestShouldExcludeContainer(t *testing.T) {
tests := []struct {
name string
@@ -1440,59 +1203,3 @@ func TestShouldExcludeContainer(t *testing.T) {
})
}
}
func TestAnsiEscapePattern(t *testing.T) {
tests := []struct {
name string
input string
expected string
}{
{
name: "no ANSI codes",
input: "Hello, World!",
expected: "Hello, World!",
},
{
name: "simple color code",
input: "\x1b[34mINFO\x1b[0m client mode",
expected: "INFO client mode",
},
{
name: "multiple color codes",
input: "\x1b[31mERROR\x1b[0m: \x1b[33mWarning\x1b[0m message",
expected: "ERROR: Warning message",
},
{
name: "bold and color",
input: "\x1b[1;32mSUCCESS\x1b[0m",
expected: "SUCCESS",
},
{
name: "cursor movement codes",
input: "Line 1\x1b[KLine 2",
expected: "Line 1Line 2",
},
{
name: "256 color code",
input: "\x1b[38;5;196mRed text\x1b[0m",
expected: "Red text",
},
{
name: "RGB/truecolor code",
input: "\x1b[38;2;255;0;0mRed text\x1b[0m",
expected: "Red text",
},
{
name: "mixed content with newlines",
input: "\x1b[34m2024-01-01 12:00:00\x1b[0m INFO Starting\n\x1b[31m2024-01-01 12:00:01\x1b[0m ERROR Failed",
expected: "2024-01-01 12:00:00 INFO Starting\n2024-01-01 12:00:01 ERROR Failed",
},
}
for _, tt := range tests {
t.Run(tt.name, func(t *testing.T) {
result := ansiEscapePattern.ReplaceAllString(tt.input, "")
assert.Equal(t, tt.expected, result)
})
}
}

View File

@@ -1,95 +0,0 @@
package agent
import (
"fmt"
"strconv"
"strings"
)
func isEmmcBlockName(name string) bool {
if !strings.HasPrefix(name, "mmcblk") {
return false
}
suffix := strings.TrimPrefix(name, "mmcblk")
if suffix == "" {
return false
}
for _, c := range suffix {
if c < '0' || c > '9' {
return false
}
}
return true
}
func parseHexOrDecByte(s string) (uint8, bool) {
s = strings.TrimSpace(s)
if s == "" {
return 0, false
}
base := 10
if strings.HasPrefix(s, "0x") || strings.HasPrefix(s, "0X") {
base = 16
s = s[2:]
}
parsed, err := strconv.ParseUint(s, base, 8)
if err != nil {
return 0, false
}
return uint8(parsed), true
}
func parseHexBytePair(s string) (uint8, uint8, bool) {
fields := strings.Fields(s)
if len(fields) < 2 {
return 0, 0, false
}
a, okA := parseHexOrDecByte(fields[0])
b, okB := parseHexOrDecByte(fields[1])
if !okA && !okB {
return 0, 0, false
}
return a, b, true
}
func emmcSmartStatus(preEOL uint8) string {
switch preEOL {
case 0x01:
return "PASSED"
case 0x02:
return "WARNING"
case 0x03:
return "FAILED"
default:
return "UNKNOWN"
}
}
func emmcPreEOLString(preEOL uint8) string {
switch preEOL {
case 0x01:
return "0x01 (normal)"
case 0x02:
return "0x02 (warning)"
case 0x03:
return "0x03 (urgent)"
default:
return fmt.Sprintf("0x%02x", preEOL)
}
}
func emmcLifeTimeString(v uint8) string {
// JEDEC eMMC: 0x01..0x0A => 0-100% used in 10% steps, 0x0B => exceeded.
switch {
case v == 0:
return "0x00 (not reported)"
case v >= 0x01 && v <= 0x0A:
low := int(v-1) * 10
high := int(v) * 10
return fmt.Sprintf("0x%02x (%d-%d%% used)", v, low, high)
case v == 0x0B:
return "0x0b (>100% used)"
default:
return fmt.Sprintf("0x%02x", v)
}
}

View File

@@ -1,78 +0,0 @@
package agent
import "testing"
func TestParseHexOrDecByte(t *testing.T) {
tests := []struct {
in string
want uint8
ok bool
}{
{"0x01", 1, true},
{"0X0b", 11, true},
{"01", 1, true},
{" 3 ", 3, true},
{"", 0, false},
{"0x", 0, false},
{"nope", 0, false},
}
for _, tt := range tests {
got, ok := parseHexOrDecByte(tt.in)
if ok != tt.ok || got != tt.want {
t.Fatalf("parseHexOrDecByte(%q) = (%d,%v), want (%d,%v)", tt.in, got, ok, tt.want, tt.ok)
}
}
}
func TestParseHexBytePair(t *testing.T) {
a, b, ok := parseHexBytePair("0x01 0x02\n")
if !ok || a != 1 || b != 2 {
t.Fatalf("parseHexBytePair hex = (%d,%d,%v), want (1,2,true)", a, b, ok)
}
a, b, ok = parseHexBytePair("01 02")
if !ok || a != 1 || b != 2 {
t.Fatalf("parseHexBytePair dec = (%d,%d,%v), want (1,2,true)", a, b, ok)
}
_, _, ok = parseHexBytePair("0x01")
if ok {
t.Fatalf("parseHexBytePair short input ok=true, want false")
}
}
func TestEmmcSmartStatus(t *testing.T) {
if got := emmcSmartStatus(0x01); got != "PASSED" {
t.Fatalf("emmcSmartStatus(0x01) = %q, want PASSED", got)
}
if got := emmcSmartStatus(0x02); got != "WARNING" {
t.Fatalf("emmcSmartStatus(0x02) = %q, want WARNING", got)
}
if got := emmcSmartStatus(0x03); got != "FAILED" {
t.Fatalf("emmcSmartStatus(0x03) = %q, want FAILED", got)
}
if got := emmcSmartStatus(0x00); got != "UNKNOWN" {
t.Fatalf("emmcSmartStatus(0x00) = %q, want UNKNOWN", got)
}
}
func TestIsEmmcBlockName(t *testing.T) {
cases := []struct {
name string
ok bool
}{
{"mmcblk0", true},
{"mmcblk1", true},
{"mmcblk10", true},
{"mmcblk0p1", false},
{"sda", false},
{"mmcblk", false},
{"mmcblkA", false},
}
for _, c := range cases {
if got := isEmmcBlockName(c.name); got != c.ok {
t.Fatalf("isEmmcBlockName(%q) = %v, want %v", c.name, got, c.ok)
}
}
}

View File

@@ -1,227 +0,0 @@
//go:build linux
package agent
import (
"os"
"path/filepath"
"strconv"
"strings"
"github.com/henrygd/beszel/internal/entities/smart"
)
// emmcSysfsRoot is a test hook; production value is "/sys".
var emmcSysfsRoot = "/sys"
type emmcHealth struct {
model string
serial string
revision string
capacity uint64
preEOL uint8
lifeA uint8
lifeB uint8
}
func scanEmmcDevices() []*DeviceInfo {
blockDir := filepath.Join(emmcSysfsRoot, "class", "block")
entries, err := os.ReadDir(blockDir)
if err != nil {
return nil
}
devices := make([]*DeviceInfo, 0, 2)
for _, ent := range entries {
name := ent.Name()
if !isEmmcBlockName(name) {
continue
}
deviceDir := filepath.Join(blockDir, name, "device")
if !hasEmmcHealthFiles(deviceDir) {
continue
}
devPath := filepath.Join("/dev", name)
devices = append(devices, &DeviceInfo{
Name: devPath,
Type: "emmc",
InfoName: devPath + " [eMMC]",
Protocol: "MMC",
})
}
return devices
}
func (sm *SmartManager) collectEmmcHealth(deviceInfo *DeviceInfo) (bool, error) {
if deviceInfo == nil || deviceInfo.Name == "" {
return false, nil
}
base := filepath.Base(deviceInfo.Name)
if !isEmmcBlockName(base) && !strings.EqualFold(deviceInfo.Type, "emmc") && !strings.EqualFold(deviceInfo.Type, "mmc") {
return false, nil
}
health, ok := readEmmcHealth(base)
if !ok {
return false, nil
}
// Normalize the device type to keep pruning logic stable across refreshes.
deviceInfo.Type = "emmc"
key := health.serial
if key == "" {
key = filepath.Join("/dev", base)
}
status := emmcSmartStatus(health.preEOL)
attrs := []*smart.SmartAttribute{
{
Name: "PreEOLInfo",
RawValue: uint64(health.preEOL),
RawString: emmcPreEOLString(health.preEOL),
},
{
Name: "DeviceLifeTimeEstA",
RawValue: uint64(health.lifeA),
RawString: emmcLifeTimeString(health.lifeA),
},
{
Name: "DeviceLifeTimeEstB",
RawValue: uint64(health.lifeB),
RawString: emmcLifeTimeString(health.lifeB),
},
}
sm.Lock()
defer sm.Unlock()
if _, exists := sm.SmartDataMap[key]; !exists {
sm.SmartDataMap[key] = &smart.SmartData{}
}
data := sm.SmartDataMap[key]
data.ModelName = health.model
data.SerialNumber = health.serial
data.FirmwareVersion = health.revision
data.Capacity = health.capacity
data.Temperature = 0
data.SmartStatus = status
data.DiskName = filepath.Join("/dev", base)
data.DiskType = "emmc"
data.Attributes = attrs
return true, nil
}
func readEmmcHealth(blockName string) (emmcHealth, bool) {
var out emmcHealth
if !isEmmcBlockName(blockName) {
return out, false
}
deviceDir := filepath.Join(emmcSysfsRoot, "class", "block", blockName, "device")
preEOL, okPre := readHexByteFile(filepath.Join(deviceDir, "pre_eol_info"))
// Some kernels expose EXT_CSD lifetime via "life_time" (two bytes), others as
// separate files. Support both.
lifeA, lifeB, okLife := readLifeTime(deviceDir)
if !okPre && !okLife {
return out, false
}
out.preEOL = preEOL
out.lifeA = lifeA
out.lifeB = lifeB
out.model = readStringFile(filepath.Join(deviceDir, "name"))
out.serial = readStringFile(filepath.Join(deviceDir, "serial"))
out.revision = readStringFile(filepath.Join(deviceDir, "prv"))
if capBytes, ok := readBlockCapacityBytes(blockName); ok {
out.capacity = capBytes
}
return out, true
}
func readLifeTime(deviceDir string) (uint8, uint8, bool) {
if content, ok := readStringFileOK(filepath.Join(deviceDir, "life_time")); ok {
a, b, ok := parseHexBytePair(content)
return a, b, ok
}
a, okA := readHexByteFile(filepath.Join(deviceDir, "device_life_time_est_typ_a"))
b, okB := readHexByteFile(filepath.Join(deviceDir, "device_life_time_est_typ_b"))
if okA || okB {
return a, b, true
}
return 0, 0, false
}
func readBlockCapacityBytes(blockName string) (uint64, bool) {
sizePath := filepath.Join(emmcSysfsRoot, "class", "block", blockName, "size")
lbsPath := filepath.Join(emmcSysfsRoot, "class", "block", blockName, "queue", "logical_block_size")
sizeStr, ok := readStringFileOK(sizePath)
if !ok {
return 0, false
}
sectors, err := strconv.ParseUint(sizeStr, 10, 64)
if err != nil || sectors == 0 {
return 0, false
}
lbsStr, ok := readStringFileOK(lbsPath)
logicalBlockSize := uint64(512)
if ok {
if parsed, err := strconv.ParseUint(lbsStr, 10, 64); err == nil && parsed > 0 {
logicalBlockSize = parsed
}
}
return sectors * logicalBlockSize, true
}
func readHexByteFile(path string) (uint8, bool) {
content, ok := readStringFileOK(path)
if !ok {
return 0, false
}
b, ok := parseHexOrDecByte(content)
return b, ok
}
func readStringFile(path string) string {
content, _ := readStringFileOK(path)
return content
}
func readStringFileOK(path string) (string, bool) {
b, err := os.ReadFile(path)
if err != nil {
return "", false
}
return strings.TrimSpace(string(b)), true
}
func hasEmmcHealthFiles(deviceDir string) bool {
entries, err := os.ReadDir(deviceDir)
if err != nil {
return false
}
for _, ent := range entries {
switch ent.Name() {
case "pre_eol_info", "life_time", "device_life_time_est_typ_a", "device_life_time_est_typ_b":
return true
}
}
return false
}

View File

@@ -1,80 +0,0 @@
//go:build linux
package agent
import (
"os"
"path/filepath"
"testing"
"github.com/henrygd/beszel/internal/entities/smart"
)
func TestEmmcMockSysfsScanAndCollect(t *testing.T) {
tmp := t.TempDir()
prev := emmcSysfsRoot
emmcSysfsRoot = tmp
t.Cleanup(func() { emmcSysfsRoot = prev })
// Fake: /sys/class/block/mmcblk0
mmcDeviceDir := filepath.Join(tmp, "class", "block", "mmcblk0", "device")
mmcQueueDir := filepath.Join(tmp, "class", "block", "mmcblk0", "queue")
if err := os.MkdirAll(mmcDeviceDir, 0o755); err != nil {
t.Fatal(err)
}
if err := os.MkdirAll(mmcQueueDir, 0o755); err != nil {
t.Fatal(err)
}
write := func(path, content string) {
t.Helper()
if err := os.WriteFile(path, []byte(content), 0o644); err != nil {
t.Fatal(err)
}
}
write(filepath.Join(mmcDeviceDir, "pre_eol_info"), "0x02\n")
write(filepath.Join(mmcDeviceDir, "life_time"), "0x04 0x05\n")
write(filepath.Join(mmcDeviceDir, "name"), "H26M52103FMR\n")
write(filepath.Join(mmcDeviceDir, "serial"), "01234567\n")
write(filepath.Join(mmcDeviceDir, "prv"), "0x08\n")
write(filepath.Join(mmcQueueDir, "logical_block_size"), "512\n")
write(filepath.Join(tmp, "class", "block", "mmcblk0", "size"), "1024\n") // sectors
devs := scanEmmcDevices()
if len(devs) != 1 {
t.Fatalf("scanEmmcDevices() = %d devices, want 1", len(devs))
}
if devs[0].Name != "/dev/mmcblk0" || devs[0].Type != "emmc" {
t.Fatalf("scanEmmcDevices()[0] = %+v, want Name=/dev/mmcblk0 Type=emmc", devs[0])
}
sm := &SmartManager{SmartDataMap: map[string]*smart.SmartData{}}
ok, err := sm.collectEmmcHealth(devs[0])
if err != nil || !ok {
t.Fatalf("collectEmmcHealth() = (ok=%v, err=%v), want (true,nil)", ok, err)
}
if len(sm.SmartDataMap) != 1 {
t.Fatalf("SmartDataMap len=%d, want 1", len(sm.SmartDataMap))
}
var got *smart.SmartData
for _, v := range sm.SmartDataMap {
got = v
break
}
if got == nil {
t.Fatalf("SmartDataMap value nil")
}
if got.DiskType != "emmc" || got.DiskName != "/dev/mmcblk0" {
t.Fatalf("disk fields = (type=%q name=%q), want (emmc,/dev/mmcblk0)", got.DiskType, got.DiskName)
}
if got.SmartStatus != "WARNING" {
t.Fatalf("SmartStatus=%q, want WARNING", got.SmartStatus)
}
if got.SerialNumber != "01234567" || got.ModelName == "" || got.Capacity == 0 {
t.Fatalf("identity fields = (model=%q serial=%q cap=%d), want non-empty model, serial 01234567, cap>0", got.ModelName, got.SerialNumber, got.Capacity)
}
if len(got.Attributes) < 3 {
t.Fatalf("attributes len=%d, want >= 3", len(got.Attributes))
}
}

View File

@@ -1,14 +0,0 @@
//go:build !linux
package agent
// Non-Linux builds: eMMC health via sysfs is not available.
func scanEmmcDevices() []*DeviceInfo {
return nil
}
func (sm *SmartManager) collectEmmcHealth(deviceInfo *DeviceInfo) (bool, error) {
return false, nil
}

View File

@@ -1,87 +0,0 @@
package agent
import (
"crypto/sha256"
"encoding/hex"
"errors"
"os"
"path/filepath"
"strings"
"github.com/shirou/gopsutil/v4/cpu"
"github.com/shirou/gopsutil/v4/host"
)
const fingerprintFileName = "fingerprint"
// knownBadUUID is a commonly known "product_uuid" that is not unique across systems.
const knownBadUUID = "03000200-0400-0500-0006-000700080009"
// GetFingerprint returns the agent fingerprint. It first tries to read a saved
// fingerprint from the data directory. If not found (or dataDir is empty), it
// generates one from system properties. The hostname and cpuModel parameters are
// used as fallback material if host.HostID() fails. If either is empty, they
// are fetched from the system automatically.
//
// If a new fingerprint is generated and a dataDir is provided, it is saved.
func GetFingerprint(dataDir, hostname, cpuModel string) string {
if dataDir != "" {
if fp, err := readFingerprint(dataDir); err == nil {
return fp
}
}
fp := generateFingerprint(hostname, cpuModel)
if dataDir != "" {
_ = SaveFingerprint(dataDir, fp)
}
return fp
}
// generateFingerprint creates a fingerprint from system properties.
// It tries host.HostID() first, falling back to hostname + cpuModel.
// If hostname or cpuModel are empty, they are fetched from the system.
func generateFingerprint(hostname, cpuModel string) string {
fingerprint, err := host.HostID()
if err != nil || fingerprint == "" || fingerprint == knownBadUUID {
if hostname == "" {
hostname, _ = os.Hostname()
}
if cpuModel == "" {
if info, err := cpu.Info(); err == nil && len(info) > 0 {
cpuModel = info[0].ModelName
}
}
fingerprint = hostname + cpuModel
}
sum := sha256.Sum256([]byte(fingerprint))
return hex.EncodeToString(sum[:24])
}
// readFingerprint reads the saved fingerprint from the data directory.
func readFingerprint(dataDir string) (string, error) {
fp, err := os.ReadFile(filepath.Join(dataDir, fingerprintFileName))
if err != nil {
return "", err
}
s := strings.TrimSpace(string(fp))
if s == "" {
return "", errors.New("fingerprint file is empty")
}
return s, nil
}
// SaveFingerprint writes the fingerprint to the data directory.
func SaveFingerprint(dataDir, fingerprint string) error {
return os.WriteFile(filepath.Join(dataDir, fingerprintFileName), []byte(fingerprint), 0o644)
}
// DeleteFingerprint removes the saved fingerprint file from the data directory.
// Returns nil if the file does not exist (idempotent).
func DeleteFingerprint(dataDir string) error {
err := os.Remove(filepath.Join(dataDir, fingerprintFileName))
if errors.Is(err, os.ErrNotExist) {
return nil
}
return err
}

View File

@@ -1,102 +0,0 @@
//go:build testing
package agent
import (
"os"
"path/filepath"
"testing"
"github.com/stretchr/testify/assert"
"github.com/stretchr/testify/require"
)
func TestGetFingerprint(t *testing.T) {
t.Run("reads existing fingerprint from file", func(t *testing.T) {
dir := t.TempDir()
expected := "abc123def456"
err := os.WriteFile(filepath.Join(dir, fingerprintFileName), []byte(expected), 0644)
require.NoError(t, err)
fp := GetFingerprint(dir, "", "")
assert.Equal(t, expected, fp)
})
t.Run("trims whitespace from file", func(t *testing.T) {
dir := t.TempDir()
err := os.WriteFile(filepath.Join(dir, fingerprintFileName), []byte(" abc123 \n"), 0644)
require.NoError(t, err)
fp := GetFingerprint(dir, "", "")
assert.Equal(t, "abc123", fp)
})
t.Run("generates fingerprint when file does not exist", func(t *testing.T) {
dir := t.TempDir()
fp := GetFingerprint(dir, "", "")
assert.NotEmpty(t, fp)
})
t.Run("generates fingerprint when dataDir is empty", func(t *testing.T) {
fp := GetFingerprint("", "", "")
assert.NotEmpty(t, fp)
})
t.Run("generates consistent fingerprint for same inputs", func(t *testing.T) {
fp1 := GetFingerprint("", "myhost", "mycpu")
fp2 := GetFingerprint("", "myhost", "mycpu")
assert.Equal(t, fp1, fp2)
})
t.Run("prefers saved fingerprint over generated", func(t *testing.T) {
dir := t.TempDir()
require.NoError(t, SaveFingerprint(dir, "saved-fp"))
fp := GetFingerprint(dir, "anyhost", "anycpu")
assert.Equal(t, "saved-fp", fp)
})
}
func TestSaveFingerprint(t *testing.T) {
t.Run("saves fingerprint to file", func(t *testing.T) {
dir := t.TempDir()
err := SaveFingerprint(dir, "abc123")
require.NoError(t, err)
content, err := os.ReadFile(filepath.Join(dir, fingerprintFileName))
require.NoError(t, err)
assert.Equal(t, "abc123", string(content))
})
t.Run("overwrites existing fingerprint", func(t *testing.T) {
dir := t.TempDir()
require.NoError(t, SaveFingerprint(dir, "old"))
require.NoError(t, SaveFingerprint(dir, "new"))
content, err := os.ReadFile(filepath.Join(dir, fingerprintFileName))
require.NoError(t, err)
assert.Equal(t, "new", string(content))
})
}
func TestDeleteFingerprint(t *testing.T) {
t.Run("deletes existing fingerprint", func(t *testing.T) {
dir := t.TempDir()
fp := filepath.Join(dir, fingerprintFileName)
err := os.WriteFile(fp, []byte("abc123"), 0644)
require.NoError(t, err)
err = DeleteFingerprint(dir)
require.NoError(t, err)
// Verify file is gone
_, err = os.Stat(fp)
assert.True(t, os.IsNotExist(err))
})
t.Run("no error when file does not exist", func(t *testing.T) {
dir := t.TempDir()
err := DeleteFingerprint(dir)
assert.NoError(t, err)
})
}

View File

@@ -5,29 +5,29 @@ import (
"bytes"
"encoding/json"
"fmt"
"log/slog"
"maps"
"os/exec"
"regexp"
"runtime"
"strconv"
"strings"
"sync"
"time"
"github.com/henrygd/beszel/internal/entities/system"
"golang.org/x/exp/slog"
)
const (
// Commands
nvidiaSmiCmd string = "nvidia-smi"
rocmSmiCmd string = "rocm-smi"
tegraStatsCmd string = "tegrastats"
nvtopCmd string = "nvtop"
powermetricsCmd string = "powermetrics"
macmonCmd string = "macmon"
noGPUFoundMsg string = "no GPU found - see https://beszel.dev/guide/gpu"
nvidiaSmiCmd string = "nvidia-smi"
rocmSmiCmd string = "rocm-smi"
tegraStatsCmd string = "tegrastats"
// Polling intervals
nvidiaSmiInterval string = "4" // in seconds
tegraStatsInterval string = "3700" // in milliseconds
rocmSmiInterval time.Duration = 4300 * time.Millisecond
// Command retry and timeout constants
retryWaitTime time.Duration = 5 * time.Second
maxFailureRetries int = 5
@@ -40,7 +40,11 @@ const (
// GPUManager manages data collection for GPUs (either Nvidia or AMD)
type GPUManager struct {
sync.Mutex
GpuDataMap map[string]*system.GPUData
nvidiaSmi bool
rocmSmi bool
tegrastats bool
intelGpuStats bool
GpuDataMap map[string]*system.GPUData
// lastAvgData stores the last calculated averages for each GPU
// Used when a collection happens before new data arrives (Count == 0)
lastAvgData map[string]system.GPUData
@@ -81,58 +85,6 @@ type gpuCollector struct {
var errNoValidData = fmt.Errorf("no valid GPU data found") // Error for missing data
// collectorSource identifies a selectable GPU collector in GPU_COLLECTOR.
type collectorSource string
const (
collectorSourceNVTop collectorSource = collectorSource(nvtopCmd)
collectorSourceNVML collectorSource = "nvml"
collectorSourceNvidiaSMI collectorSource = collectorSource(nvidiaSmiCmd)
collectorSourceIntelGpuTop collectorSource = collectorSource(intelGpuStatsCmd)
collectorSourceAmdSysfs collectorSource = "amd_sysfs"
collectorSourceRocmSMI collectorSource = collectorSource(rocmSmiCmd)
collectorSourceMacmon collectorSource = collectorSource(macmonCmd)
collectorSourcePowermetrics collectorSource = collectorSource(powermetricsCmd)
collectorGroupNvidia string = "nvidia"
collectorGroupIntel string = "intel"
collectorGroupAmd string = "amd"
collectorGroupApple string = "apple"
)
func isValidCollectorSource(source collectorSource) bool {
switch source {
case collectorSourceNVTop,
collectorSourceNVML,
collectorSourceNvidiaSMI,
collectorSourceIntelGpuTop,
collectorSourceAmdSysfs,
collectorSourceRocmSMI,
collectorSourceMacmon,
collectorSourcePowermetrics:
return true
}
return false
}
// gpuCapabilities describes detected GPU tooling and sysfs support on the host.
type gpuCapabilities struct {
hasNvidiaSmi bool
hasRocmSmi bool
hasAmdSysfs bool
hasTegrastats bool
hasIntelGpuTop bool
hasNvtop bool
hasMacmon bool
hasPowermetrics bool
}
type collectorDefinition struct {
group string
available bool
start func(onFailure func()) bool
deprecationWarning string
}
// starts and manages the ongoing collection of GPU data for the specified GPU management utility
func (c *gpuCollector) start() {
for {
@@ -184,10 +136,10 @@ func (gm *GPUManager) getJetsonParser() func(output []byte) bool {
// use closure to avoid recompiling the regex
ramPattern := regexp.MustCompile(`RAM (\d+)/(\d+)MB`)
gr3dPattern := regexp.MustCompile(`GR3D_FREQ (\d+)%`)
tempPattern := regexp.MustCompile(`(?:tj|GPU)@(\d+\.?\d*)C`)
tempPattern := regexp.MustCompile(`tj@(\d+\.?\d*)C`)
// Orin Nano / NX do not have GPU specific power monitor
// TODO: Maybe use VDD_IN for Nano / NX and add a total system power chart
powerPattern := regexp.MustCompile(`(GPU_SOC|CPU_GPU_CV)\s+(\d+)mW|VDD_SYS_GPU\s+(\d+)/\d+`)
powerPattern := regexp.MustCompile(`(GPU_SOC|CPU_GPU_CV) (\d+)mW`)
// jetson devices have only one gpu so we'll just initialize here
gpuData := &system.GPUData{Name: "GPU"}
@@ -216,13 +168,7 @@ func (gm *GPUManager) getJetsonParser() func(output []byte) bool {
// Parse power usage
powerMatches := powerPattern.FindSubmatch(output)
if powerMatches != nil {
// powerMatches[2] is the "(GPU_SOC|CPU_GPU_CV) <N>mW" capture
// powerMatches[3] is the "VDD_SYS_GPU <N>/<N>" capture
powerStr := string(powerMatches[2])
if powerStr == "" {
powerStr = string(powerMatches[3])
}
power, _ := strconv.ParseFloat(powerStr, 64)
power, _ := strconv.ParseFloat(string(powerMatches[2]), 64)
gpuData.Power += power / milliwattsInAWatt
}
gpuData.Count++
@@ -285,11 +231,10 @@ func (gm *GPUManager) parseAmdData(output []byte) bool {
totalMemory, _ := strconv.ParseFloat(v.MemoryTotal, 64)
usage, _ := strconv.ParseFloat(v.Usage, 64)
id := v.ID
if _, ok := gm.GpuDataMap[id]; !ok {
gm.GpuDataMap[id] = &system.GPUData{Name: v.Name}
if _, ok := gm.GpuDataMap[v.ID]; !ok {
gm.GpuDataMap[v.ID] = &system.GPUData{Name: v.Name}
}
gpu := gm.GpuDataMap[id]
gpu := gm.GpuDataMap[v.ID]
gpu.Temperature, _ = strconv.ParseFloat(v.Temperature, 64)
gpu.MemoryUsed = bytesToMegabytes(memoryUsage)
gpu.MemoryTotal = bytesToMegabytes(totalMemory)
@@ -352,13 +297,8 @@ func (gm *GPUManager) calculateGPUAverage(id string, gpu *system.GPUData, cacheK
currentCount := uint32(gpu.Count)
deltaCount := gm.calculateDeltaCount(currentCount, lastSnapshot)
// If no new data arrived
// If no new data arrived, use last known average
if deltaCount == 0 {
// If GPU appears suspended (instantaneous values are 0), return zero values
// Otherwise return last known average for temporary collection gaps
if gpu.Temperature == 0 && gpu.MemoryUsed == 0 {
return system.GPUData{Name: gpu.Name}
}
return gm.lastAvgData[id] // zero value if not found
}
@@ -438,324 +378,105 @@ func (gm *GPUManager) storeSnapshot(id string, gpu *system.GPUData, cacheKey uin
gm.lastSnapshots[cacheKey][id] = snapshot
}
// discoverGpuCapabilities checks for available GPU tooling and sysfs support.
// It only reports capability presence and does not apply policy decisions.
func (gm *GPUManager) discoverGpuCapabilities() gpuCapabilities {
caps := gpuCapabilities{
hasAmdSysfs: gm.hasAmdSysfs(),
}
// detectGPUs checks for the presence of GPU management tools (nvidia-smi, rocm-smi, tegrastats)
// in the system path. It sets the corresponding flags in the GPUManager struct if any of these
// tools are found. If none of the tools are found, it returns an error indicating that no GPU
// management tools are available.
func (gm *GPUManager) detectGPUs() error {
if _, err := exec.LookPath(nvidiaSmiCmd); err == nil {
caps.hasNvidiaSmi = true
gm.nvidiaSmi = true
}
if _, err := exec.LookPath(rocmSmiCmd); err == nil {
caps.hasRocmSmi = true
gm.rocmSmi = true
}
if _, err := exec.LookPath(tegraStatsCmd); err == nil {
caps.hasTegrastats = true
gm.tegrastats = true
gm.nvidiaSmi = false
}
if _, err := exec.LookPath(intelGpuStatsCmd); err == nil {
caps.hasIntelGpuTop = true
gm.intelGpuStats = true
}
if _, err := exec.LookPath(nvtopCmd); err == nil {
caps.hasNvtop = true
if gm.nvidiaSmi || gm.rocmSmi || gm.tegrastats || gm.intelGpuStats {
return nil
}
if runtime.GOOS == "darwin" {
if _, err := exec.LookPath(macmonCmd); err == nil {
caps.hasMacmon = true
}
if _, err := exec.LookPath(powermetricsCmd); err == nil {
caps.hasPowermetrics = true
}
}
return caps
return fmt.Errorf("no GPU found - install nvidia-smi, rocm-smi, tegrastats, or intel_gpu_top")
}
func hasAnyGpuCollector(caps gpuCapabilities) bool {
return caps.hasNvidiaSmi || caps.hasRocmSmi || caps.hasAmdSysfs || caps.hasTegrastats || caps.hasIntelGpuTop || caps.hasNvtop || caps.hasMacmon || caps.hasPowermetrics
}
func (gm *GPUManager) startIntelCollector() {
go func() {
failures := 0
for {
if err := gm.collectIntelStats(); err != nil {
failures++
if failures > maxFailureRetries {
break
}
slog.Warn("Error collecting Intel GPU data; see https://beszel.dev/guide/gpu", "err", err)
time.Sleep(retryWaitTime)
continue
}
}
}()
}
func (gm *GPUManager) startNvidiaSmiCollector(intervalSeconds string) {
// startCollector starts the appropriate GPU data collector based on the command
func (gm *GPUManager) startCollector(command string) {
collector := gpuCollector{
name: nvidiaSmiCmd,
name: command,
bufSize: 10 * 1024,
cmdArgs: []string{
"-l", intervalSeconds,
}
switch command {
case intelGpuStatsCmd:
go func() {
failures := 0
for {
if err := gm.collectIntelStats(); err != nil {
failures++
if failures > maxFailureRetries {
break
}
slog.Warn("Error collecting Intel GPU data; see https://beszel.dev/guide/gpu", "err", err)
time.Sleep(retryWaitTime)
continue
}
}
}()
case nvidiaSmiCmd:
collector.cmdArgs = []string{
"-l", nvidiaSmiInterval,
"--query-gpu=index,name,temperature.gpu,memory.used,memory.total,utilization.gpu,power.draw",
"--format=csv,noheader,nounits",
},
parse: gm.parseNvidiaData,
}
go collector.start()
}
func (gm *GPUManager) startTegraStatsCollector(intervalMilliseconds string) {
collector := gpuCollector{
name: tegraStatsCmd,
bufSize: 10 * 1024,
cmdArgs: []string{"--interval", intervalMilliseconds},
parse: gm.getJetsonParser(),
}
go collector.start()
}
func (gm *GPUManager) startRocmSmiCollector(pollInterval time.Duration) {
collector := gpuCollector{
name: rocmSmiCmd,
bufSize: 10 * 1024,
cmdArgs: []string{"--showid", "--showtemp", "--showuse", "--showpower", "--showproductname", "--showmeminfo", "vram", "--json"},
parse: gm.parseAmdData,
}
go func() {
failures := 0
for {
if err := collector.collect(); err != nil {
failures++
if failures > maxFailureRetries {
break
}
collector.parse = gm.parseNvidiaData
go collector.start()
case tegraStatsCmd:
collector.cmdArgs = []string{"--interval", tegraStatsInterval}
collector.parse = gm.getJetsonParser()
go collector.start()
case rocmSmiCmd:
collector.cmdArgs = []string{"--showid", "--showtemp", "--showuse", "--showpower", "--showproductname", "--showmeminfo", "vram", "--json"}
collector.parse = gm.parseAmdData
go func() {
failures := 0
for {
if err := collector.collect(); err != nil {
failures++
if failures > maxFailureRetries {
break
}
slog.Warn("Error collecting AMD GPU data", "err", err)
}
slog.Warn("Error collecting AMD GPU data via rocm-smi", "err", err)
time.Sleep(rocmSmiInterval)
}
time.Sleep(pollInterval)
}
}()
}
func (gm *GPUManager) collectorDefinitions(caps gpuCapabilities) map[collectorSource]collectorDefinition {
return map[collectorSource]collectorDefinition{
collectorSourceNVML: {
group: collectorGroupNvidia,
available: caps.hasNvidiaSmi,
start: func(_ func()) bool {
return gm.startNvmlCollector()
},
},
collectorSourceNvidiaSMI: {
group: collectorGroupNvidia,
available: caps.hasNvidiaSmi,
start: func(_ func()) bool {
gm.startNvidiaSmiCollector("4") // seconds
return true
},
},
collectorSourceIntelGpuTop: {
group: collectorGroupIntel,
available: caps.hasIntelGpuTop,
start: func(_ func()) bool {
gm.startIntelCollector()
return true
},
},
collectorSourceAmdSysfs: {
group: collectorGroupAmd,
available: caps.hasAmdSysfs,
start: func(_ func()) bool {
return gm.startAmdSysfsCollector()
},
},
collectorSourceRocmSMI: {
group: collectorGroupAmd,
available: caps.hasRocmSmi,
deprecationWarning: "rocm-smi is deprecated and may be removed in a future release",
start: func(_ func()) bool {
gm.startRocmSmiCollector(4300 * time.Millisecond)
return true
},
},
collectorSourceNVTop: {
available: caps.hasNvtop,
start: func(onFailure func()) bool {
gm.startNvtopCollector("30", onFailure) // tens of milliseconds
return true
},
},
collectorSourceMacmon: {
group: collectorGroupApple,
available: caps.hasMacmon,
start: func(_ func()) bool {
gm.startMacmonCollector()
return true
},
},
collectorSourcePowermetrics: {
group: collectorGroupApple,
available: caps.hasPowermetrics,
start: func(_ func()) bool {
gm.startPowermetricsCollector()
return true
},
},
}()
}
}
// parseCollectorPriority parses GPU_COLLECTOR and returns valid ordered entries.
func parseCollectorPriority(value string) []collectorSource {
parts := strings.Split(value, ",")
priorities := make([]collectorSource, 0, len(parts))
for _, raw := range parts {
name := collectorSource(strings.TrimSpace(strings.ToLower(raw)))
if !isValidCollectorSource(name) {
if name != "" {
slog.Warn("Ignoring unknown GPU collector", "collector", name)
}
continue
}
priorities = append(priorities, name)
}
return priorities
}
// startNvmlCollector initializes NVML and starts its polling loop.
func (gm *GPUManager) startNvmlCollector() bool {
collector := &nvmlCollector{gm: gm}
if err := collector.init(); err != nil {
slog.Warn("Failed to initialize NVML", "err", err)
return false
}
go collector.start()
return true
}
// startAmdSysfsCollector starts AMD GPU collection via sysfs.
func (gm *GPUManager) startAmdSysfsCollector() bool {
go func() {
if err := gm.collectAmdStats(); err != nil {
slog.Warn("Error collecting AMD GPU data via sysfs", "err", err)
}
}()
return true
}
// startCollectorsByPriority starts collectors in order with one source per vendor group.
func (gm *GPUManager) startCollectorsByPriority(priorities []collectorSource, caps gpuCapabilities) int {
definitions := gm.collectorDefinitions(caps)
selectedGroups := make(map[string]bool, 3)
started := 0
for i, source := range priorities {
definition, ok := definitions[source]
if !ok || !definition.available {
continue
}
// nvtop is not a vendor-specific collector, so should only be used if no other collectors are selected or it is first in GPU_COLLECTOR.
if source == collectorSourceNVTop {
if len(selectedGroups) > 0 {
slog.Warn("Skipping nvtop because other collectors are selected")
continue
}
// if nvtop fails, fall back to remaining collectors.
remaining := append([]collectorSource(nil), priorities[i+1:]...)
if definition.start(func() {
gm.startCollectorsByPriority(remaining, caps)
}) {
started++
return started
}
}
group := definition.group
if group == "" || selectedGroups[group] {
continue
}
if definition.deprecationWarning != "" {
slog.Warn(definition.deprecationWarning)
}
if definition.start(nil) {
selectedGroups[group] = true
started++
}
}
return started
}
// resolveLegacyCollectorPriority builds the default collector order when GPU_COLLECTOR is unset.
func (gm *GPUManager) resolveLegacyCollectorPriority(caps gpuCapabilities) []collectorSource {
priorities := make([]collectorSource, 0, 4)
if caps.hasNvidiaSmi && !caps.hasTegrastats {
if nvml, _ := GetEnv("NVML"); nvml == "true" {
priorities = append(priorities, collectorSourceNVML, collectorSourceNvidiaSMI)
} else {
priorities = append(priorities, collectorSourceNvidiaSMI)
}
}
if caps.hasRocmSmi {
if val, _ := GetEnv("AMD_SYSFS"); val == "true" {
priorities = append(priorities, collectorSourceAmdSysfs)
} else {
priorities = append(priorities, collectorSourceRocmSMI)
}
} else if caps.hasAmdSysfs {
priorities = append(priorities, collectorSourceAmdSysfs)
}
if caps.hasIntelGpuTop {
priorities = append(priorities, collectorSourceIntelGpuTop)
}
// Apple collectors are currently opt-in only for testing.
// Enable them with GPU_COLLECTOR=macmon or GPU_COLLECTOR=powermetrics.
// TODO: uncomment below when Apple collectors are confirmed to be working.
//
// Prefer macmon on macOS (no sudo). Fall back to powermetrics if present.
// if caps.hasMacmon {
// priorities = append(priorities, collectorSourceMacmon)
// } else if caps.hasPowermetrics {
// priorities = append(priorities, collectorSourcePowermetrics)
// }
// Keep nvtop as a last resort only when no vendor collector exists.
if len(priorities) == 0 && caps.hasNvtop {
priorities = append(priorities, collectorSourceNVTop)
}
return priorities
}
// NewGPUManager creates and initializes a new GPUManager
func NewGPUManager() (*GPUManager, error) {
if skipGPU, _ := GetEnv("SKIP_GPU"); skipGPU == "true" {
return nil, nil
}
var gm GPUManager
caps := gm.discoverGpuCapabilities()
if !hasAnyGpuCollector(caps) {
return nil, fmt.Errorf(noGPUFoundMsg)
if err := gm.detectGPUs(); err != nil {
return nil, err
}
gm.GpuDataMap = make(map[string]*system.GPUData)
// Jetson devices should always use tegrastats (ignore GPU_COLLECTOR).
if caps.hasTegrastats {
gm.startTegraStatsCollector("3700")
return &gm, nil
if gm.nvidiaSmi {
gm.startCollector(nvidiaSmiCmd)
}
// if GPU_COLLECTOR is set, start user-defined collectors.
if collectorConfig, ok := GetEnv("GPU_COLLECTOR"); ok && strings.TrimSpace(collectorConfig) != "" {
priorities := parseCollectorPriority(collectorConfig)
if gm.startCollectorsByPriority(priorities, caps) == 0 {
return nil, fmt.Errorf("no configured GPU collectors are available")
}
return &gm, nil
if gm.rocmSmi {
gm.startCollector(rocmSmiCmd)
}
// auto-detect and start collectors when GPU_COLLECTOR is unset.
if gm.startCollectorsByPriority(gm.resolveLegacyCollectorPriority(caps), caps) == 0 {
return nil, fmt.Errorf(noGPUFoundMsg)
if gm.tegrastats {
gm.startCollector(tegraStatsCmd)
}
if gm.intelGpuStats {
gm.startCollector(intelGpuStatsCmd)
}
return &gm, nil

View File

@@ -1,302 +0,0 @@
//go:build linux
package agent
import (
"bufio"
"fmt"
"log/slog"
"os"
"path/filepath"
"strconv"
"strings"
"sync"
"time"
"github.com/henrygd/beszel/internal/entities/system"
)
var amdgpuNameCache = struct {
sync.RWMutex
hits map[string]string
misses map[string]struct{}
}{
hits: make(map[string]string),
misses: make(map[string]struct{}),
}
// hasAmdSysfs returns true if any AMD GPU sysfs nodes are found
func (gm *GPUManager) hasAmdSysfs() bool {
cards, err := filepath.Glob("/sys/class/drm/card*/device/vendor")
if err != nil {
return false
}
for _, vendorPath := range cards {
vendor, err := os.ReadFile(vendorPath)
if err == nil && strings.TrimSpace(string(vendor)) == "0x1002" {
return true
}
}
return false
}
// collectAmdStats collects AMD GPU metrics directly from sysfs to avoid the overhead of rocm-smi
func (gm *GPUManager) collectAmdStats() error {
sysfsPollInterval := 3000 * time.Millisecond
cards, err := filepath.Glob("/sys/class/drm/card*")
if err != nil {
return err
}
var amdGpuPaths []string
for _, card := range cards {
// Ignore symbolic links and non-main card directories
if strings.Contains(filepath.Base(card), "-") || !isAmdGpu(card) {
continue
}
amdGpuPaths = append(amdGpuPaths, card)
}
if len(amdGpuPaths) == 0 {
return errNoValidData
}
slog.Debug("Using sysfs for AMD GPU data collection")
failures := 0
for {
hasData := false
for _, cardPath := range amdGpuPaths {
if gm.updateAmdGpuData(cardPath) {
hasData = true
}
}
if !hasData {
failures++
if failures > maxFailureRetries {
return errNoValidData
}
slog.Warn("No AMD GPU data from sysfs", "failures", failures)
time.Sleep(retryWaitTime)
continue
}
failures = 0
time.Sleep(sysfsPollInterval)
}
}
// isAmdGpu checks whether a DRM card path belongs to AMD vendor ID 0x1002.
func isAmdGpu(cardPath string) bool {
vendorPath := filepath.Join(cardPath, "device/vendor")
vendor, err := os.ReadFile(vendorPath)
if err != nil {
return false
}
return strings.TrimSpace(string(vendor)) == "0x1002"
}
// updateAmdGpuData reads GPU metrics from sysfs and updates the GPU data map.
// Returns true if at least some data was successfully read.
func (gm *GPUManager) updateAmdGpuData(cardPath string) bool {
devicePath := filepath.Join(cardPath, "device")
id := filepath.Base(cardPath)
// Read all sysfs values first (no lock needed - these can be slow)
usage, usageErr := readSysfsFloat(filepath.Join(devicePath, "gpu_busy_percent"))
memUsed, memUsedErr := readSysfsFloat(filepath.Join(devicePath, "mem_info_vram_used"))
memTotal, _ := readSysfsFloat(filepath.Join(devicePath, "mem_info_vram_total"))
// if gtt is present, add it to the memory used and total (https://github.com/henrygd/beszel/issues/1569#issuecomment-3837640484)
if gttUsed, err := readSysfsFloat(filepath.Join(devicePath, "mem_info_gtt_used")); err == nil && gttUsed > 0 {
if gttTotal, err := readSysfsFloat(filepath.Join(devicePath, "mem_info_gtt_total")); err == nil {
memUsed += gttUsed
memTotal += gttTotal
}
}
var temp, power float64
hwmons, _ := filepath.Glob(filepath.Join(devicePath, "hwmon/hwmon*"))
for _, hwmonDir := range hwmons {
if t, err := readSysfsFloat(filepath.Join(hwmonDir, "temp1_input")); err == nil {
temp = t / 1000.0
}
if p, err := readSysfsFloat(filepath.Join(hwmonDir, "power1_average")); err == nil {
power += p / 1000000.0
} else if p, err := readSysfsFloat(filepath.Join(hwmonDir, "power1_input")); err == nil {
power += p / 1000000.0
}
}
// Check if we got any meaningful data
if usageErr != nil && memUsedErr != nil && temp == 0 {
return false
}
// Single lock to update all values atomically
gm.Lock()
defer gm.Unlock()
gpu, ok := gm.GpuDataMap[id]
if !ok {
gpu = &system.GPUData{Name: getAmdGpuName(devicePath)}
gm.GpuDataMap[id] = gpu
}
if usageErr == nil {
gpu.Usage += usage
}
gpu.MemoryUsed = bytesToMegabytes(memUsed)
gpu.MemoryTotal = bytesToMegabytes(memTotal)
gpu.Temperature = temp
gpu.Power += power
gpu.Count++
return true
}
// readSysfsFloat reads and parses a numeric value from a sysfs file.
func readSysfsFloat(path string) (float64, error) {
val, err := os.ReadFile(path)
if err != nil {
return 0, err
}
return strconv.ParseFloat(strings.TrimSpace(string(val)), 64)
}
// normalizeHexID normalizes hex IDs by trimming spaces, lowercasing, and dropping 0x.
func normalizeHexID(id string) string {
return strings.TrimPrefix(strings.ToLower(strings.TrimSpace(id)), "0x")
}
// cacheKeyForAmdgpu builds the cache key for a device and optional revision.
func cacheKeyForAmdgpu(deviceID, revisionID string) string {
if revisionID != "" {
return deviceID + ":" + revisionID
}
return deviceID
}
// lookupAmdgpuNameInFile resolves an AMDGPU name from amdgpu.ids by device/revision.
func lookupAmdgpuNameInFile(deviceID, revisionID, filePath string) (name string, exact bool, found bool) {
file, err := os.Open(filePath)
if err != nil {
return "", false, false
}
defer file.Close()
var byDevice string
scanner := bufio.NewScanner(file)
for scanner.Scan() {
line := strings.TrimSpace(scanner.Text())
if line == "" || strings.HasPrefix(line, "#") {
continue
}
parts := strings.SplitN(line, ",", 3)
if len(parts) != 3 {
continue
}
dev := normalizeHexID(parts[0])
rev := normalizeHexID(parts[1])
productName := strings.TrimSpace(parts[2])
if dev == "" || productName == "" || dev != deviceID {
continue
}
if byDevice == "" {
byDevice = productName
}
if revisionID != "" && rev == revisionID {
return productName, true, true
}
}
if byDevice != "" {
return byDevice, false, true
}
return "", false, false
}
// getCachedAmdgpuName returns cached hit/miss status for the given device/revision.
func getCachedAmdgpuName(deviceID, revisionID string) (name string, found bool, done bool) {
// Build the list of cache keys to check. We always look up the exact device+revision key.
// When revisionID is set, we also look up deviceID alone, since the cache may store a
// device-only fallback when we couldn't resolve the exact revision.
keys := []string{cacheKeyForAmdgpu(deviceID, revisionID)}
if revisionID != "" {
keys = append(keys, deviceID)
}
knownMisses := 0
amdgpuNameCache.RLock()
defer amdgpuNameCache.RUnlock()
for _, key := range keys {
if name, ok := amdgpuNameCache.hits[key]; ok {
return name, true, true
}
if _, ok := amdgpuNameCache.misses[key]; ok {
knownMisses++
}
}
// done=true means "don't bother doing slow lookup": we either found a name (above) or
// every key we checked was already a known miss, so we've tried before and failed.
return "", false, knownMisses == len(keys)
}
// normalizeAmdgpuName trims standard suffixes from AMDGPU product names.
func normalizeAmdgpuName(name string) string {
for _, suffix := range []string{" Graphics", " Series"} {
name = strings.TrimSuffix(name, suffix)
}
return name
}
// cacheAmdgpuName stores a resolved AMDGPU name in the lookup cache.
func cacheAmdgpuName(deviceID, revisionID, name string, exact bool) {
name = normalizeAmdgpuName(name)
amdgpuNameCache.Lock()
defer amdgpuNameCache.Unlock()
if exact && revisionID != "" {
amdgpuNameCache.hits[cacheKeyForAmdgpu(deviceID, revisionID)] = name
}
amdgpuNameCache.hits[deviceID] = name
}
// cacheMissingAmdgpuName records unresolved device/revision lookups.
func cacheMissingAmdgpuName(deviceID, revisionID string) {
amdgpuNameCache.Lock()
defer amdgpuNameCache.Unlock()
amdgpuNameCache.misses[deviceID] = struct{}{}
if revisionID != "" {
amdgpuNameCache.misses[cacheKeyForAmdgpu(deviceID, revisionID)] = struct{}{}
}
}
// getAmdGpuName attempts to get a descriptive GPU name.
// First tries product_name (rarely available), then looks up the PCI device ID.
// Falls back to showing the raw device ID if not found in the lookup table.
func getAmdGpuName(devicePath string) string {
// Try product_name first (works for some enterprise GPUs)
if prod, err := os.ReadFile(filepath.Join(devicePath, "product_name")); err == nil {
return strings.TrimSpace(string(prod))
}
// Read PCI device ID and look it up
if deviceID, err := os.ReadFile(filepath.Join(devicePath, "device")); err == nil {
id := normalizeHexID(string(deviceID))
revision := ""
if revBytes, revErr := os.ReadFile(filepath.Join(devicePath, "revision")); revErr == nil {
revision = normalizeHexID(string(revBytes))
}
if name, found, done := getCachedAmdgpuName(id, revision); found {
return name
} else if !done {
if name, exact, ok := lookupAmdgpuNameInFile(id, revision, "/usr/share/libdrm/amdgpu.ids"); ok {
cacheAmdgpuName(id, revision, name, exact)
return normalizeAmdgpuName(name)
}
cacheMissingAmdgpuName(id, revision)
}
return fmt.Sprintf("AMD GPU (%s)", id)
}
return "AMD GPU"
}

View File

@@ -1,264 +0,0 @@
//go:build linux
package agent
import (
"os"
"path/filepath"
"testing"
"github.com/henrygd/beszel/internal/entities/system"
"github.com/stretchr/testify/assert"
"github.com/stretchr/testify/require"
)
func TestNormalizeHexID(t *testing.T) {
tests := []struct {
in string
want string
}{
{"0x1002", "1002"},
{"C2", "c2"},
{" 15BF ", "15bf"},
{"0x15bf", "15bf"},
{"", ""},
}
for _, tt := range tests {
subName := tt.in
if subName == "" {
subName = "empty_string"
}
t.Run(subName, func(t *testing.T) {
got := normalizeHexID(tt.in)
assert.Equal(t, tt.want, got)
})
}
}
func TestCacheKeyForAmdgpu(t *testing.T) {
tests := []struct {
deviceID string
revisionID string
want string
}{
{"1114", "c2", "1114:c2"},
{"15bf", "", "15bf"},
{"1506", "c1", "1506:c1"},
}
for _, tt := range tests {
got := cacheKeyForAmdgpu(tt.deviceID, tt.revisionID)
assert.Equal(t, tt.want, got)
}
}
func TestReadSysfsFloat(t *testing.T) {
dir := t.TempDir()
validPath := filepath.Join(dir, "val")
require.NoError(t, os.WriteFile(validPath, []byte(" 42.5 \n"), 0o644))
got, err := readSysfsFloat(validPath)
require.NoError(t, err)
assert.Equal(t, 42.5, got)
// Integer and scientific
sciPath := filepath.Join(dir, "sci")
require.NoError(t, os.WriteFile(sciPath, []byte("1e2"), 0o644))
got, err = readSysfsFloat(sciPath)
require.NoError(t, err)
assert.Equal(t, 100.0, got)
// Missing file
_, err = readSysfsFloat(filepath.Join(dir, "missing"))
require.Error(t, err)
// Invalid content
badPath := filepath.Join(dir, "bad")
require.NoError(t, os.WriteFile(badPath, []byte("not a number"), 0o644))
_, err = readSysfsFloat(badPath)
require.Error(t, err)
}
func TestIsAmdGpu(t *testing.T) {
dir := t.TempDir()
deviceDir := filepath.Join(dir, "device")
require.NoError(t, os.MkdirAll(deviceDir, 0o755))
// AMD vendor 0x1002 -> true
require.NoError(t, os.WriteFile(filepath.Join(deviceDir, "vendor"), []byte("0x1002\n"), 0o644))
assert.True(t, isAmdGpu(dir), "vendor 0x1002 should be AMD")
// Non-AMD vendor -> false
require.NoError(t, os.WriteFile(filepath.Join(deviceDir, "vendor"), []byte("0x10de\n"), 0o644))
assert.False(t, isAmdGpu(dir), "vendor 0x10de should not be AMD")
// Missing vendor file -> false
require.NoError(t, os.Remove(filepath.Join(deviceDir, "vendor")))
assert.False(t, isAmdGpu(dir), "missing vendor file should be false")
}
func TestAmdgpuNameCacheRoundTrip(t *testing.T) {
// Cache a name and retrieve it (unique key to avoid affecting other tests)
deviceID, revisionID := "cachedev99", "00"
cacheAmdgpuName(deviceID, revisionID, "AMD Test GPU 99 Graphics", true)
name, found, done := getCachedAmdgpuName(deviceID, revisionID)
assert.True(t, found)
assert.True(t, done)
assert.Equal(t, "AMD Test GPU 99", name)
// Device-only key also stored
name2, found2, _ := getCachedAmdgpuName(deviceID, "")
assert.True(t, found2)
assert.Equal(t, "AMD Test GPU 99", name2)
// Cache a miss
cacheMissingAmdgpuName("missedev99", "ab")
_, found3, done3 := getCachedAmdgpuName("missedev99", "ab")
assert.False(t, found3)
assert.True(t, done3, "done should be true so caller skips file lookup")
}
func TestUpdateAmdGpuDataWithFakeSysfs(t *testing.T) {
tests := []struct {
name string
writeGTT bool
wantMemoryUsed float64
wantMemoryTotal float64
}{
{
name: "sums vram and gtt when gtt is present",
writeGTT: true,
wantMemoryUsed: bytesToMegabytes(1073741824 + 536870912),
wantMemoryTotal: bytesToMegabytes(2147483648 + 4294967296),
},
{
name: "falls back to vram when gtt is missing",
writeGTT: false,
wantMemoryUsed: bytesToMegabytes(1073741824),
wantMemoryTotal: bytesToMegabytes(2147483648),
},
}
for _, tt := range tests {
t.Run(tt.name, func(t *testing.T) {
dir := t.TempDir()
cardPath := filepath.Join(dir, "card0")
devicePath := filepath.Join(cardPath, "device")
hwmonPath := filepath.Join(devicePath, "hwmon", "hwmon0")
require.NoError(t, os.MkdirAll(hwmonPath, 0o755))
write := func(name, content string) {
require.NoError(t, os.WriteFile(filepath.Join(devicePath, name), []byte(content), 0o644))
}
write("vendor", "0x1002")
write("device", "0x1506")
write("revision", "0xc1")
write("gpu_busy_percent", "25")
write("mem_info_vram_used", "1073741824")
write("mem_info_vram_total", "2147483648")
if tt.writeGTT {
write("mem_info_gtt_used", "536870912")
write("mem_info_gtt_total", "4294967296")
}
require.NoError(t, os.WriteFile(filepath.Join(hwmonPath, "temp1_input"), []byte("45000"), 0o644))
require.NoError(t, os.WriteFile(filepath.Join(hwmonPath, "power1_input"), []byte("20000000"), 0o644))
// Pre-cache name so getAmdGpuName returns a known value (it uses system amdgpu.ids path)
cacheAmdgpuName("1506", "c1", "AMD Radeon 610M Graphics", true)
gm := &GPUManager{GpuDataMap: make(map[string]*system.GPUData)}
ok := gm.updateAmdGpuData(cardPath)
require.True(t, ok)
gpu, ok := gm.GpuDataMap["card0"]
require.True(t, ok)
assert.Equal(t, "AMD Radeon 610M", gpu.Name)
assert.Equal(t, 25.0, gpu.Usage)
assert.Equal(t, tt.wantMemoryUsed, gpu.MemoryUsed)
assert.Equal(t, tt.wantMemoryTotal, gpu.MemoryTotal)
assert.Equal(t, 45.0, gpu.Temperature)
assert.Equal(t, 20.0, gpu.Power)
assert.Equal(t, 1.0, gpu.Count)
})
}
}
func TestLookupAmdgpuNameInFile(t *testing.T) {
idsPath := filepath.Join("test-data", "amdgpu.ids")
tests := []struct {
name string
deviceID string
revisionID string
wantName string
wantExact bool
wantFound bool
}{
{
name: "exact device and revision match",
deviceID: "1114",
revisionID: "c2",
wantName: "AMD Radeon 860M Graphics",
wantExact: true,
wantFound: true,
},
{
name: "exact match 15BF revision 01 returns 760M",
deviceID: "15bf",
revisionID: "01",
wantName: "AMD Radeon 760M Graphics",
wantExact: true,
wantFound: true,
},
{
name: "exact match 15BF revision 00 returns 780M",
deviceID: "15bf",
revisionID: "00",
wantName: "AMD Radeon 780M Graphics",
wantExact: true,
wantFound: true,
},
{
name: "device-only match returns first entry for device",
deviceID: "1506",
revisionID: "",
wantName: "AMD Radeon 610M",
wantExact: false,
wantFound: true,
},
{
name: "unknown device not found",
deviceID: "dead",
revisionID: "00",
wantName: "",
wantExact: false,
wantFound: false,
},
}
for _, tt := range tests {
t.Run(tt.name, func(t *testing.T) {
gotName, gotExact, gotFound := lookupAmdgpuNameInFile(tt.deviceID, tt.revisionID, idsPath)
assert.Equal(t, tt.wantName, gotName, "name")
assert.Equal(t, tt.wantExact, gotExact, "exact")
assert.Equal(t, tt.wantFound, gotFound, "found")
})
}
}
func TestGetAmdGpuNameFromIdsFile(t *testing.T) {
// Test that getAmdGpuName resolves a name when we can't inject the ids path.
// We only verify behavior when product_name is missing and device/revision
// would be read from sysfs; the actual lookup uses /usr/share/libdrm/amdgpu.ids.
// So this test focuses on normalizeAmdgpuName and that lookupAmdgpuNameInFile
// returns the expected name for our test-data file.
idsPath := filepath.Join("test-data", "amdgpu.ids")
name, exact, found := lookupAmdgpuNameInFile("1435", "ae", idsPath)
require.True(t, found)
require.True(t, exact)
assert.Equal(t, "AMD Custom GPU 0932", name)
assert.Equal(t, "AMD Custom GPU 0932", normalizeAmdgpuName(name))
// " Graphics" suffix is trimmed by normalizeAmdgpuName
name2 := "AMD Radeon 860M Graphics"
assert.Equal(t, "AMD Radeon 860M", normalizeAmdgpuName(name2))
}

View File

@@ -1,15 +0,0 @@
//go:build !linux
package agent
import (
"errors"
)
func (gm *GPUManager) hasAmdSysfs() bool {
return false
}
func (gm *GPUManager) collectAmdStats() error {
return errors.ErrUnsupported
}

View File

@@ -1,252 +0,0 @@
//go:build darwin
package agent
import (
"bufio"
"bytes"
"encoding/json"
"io"
"log/slog"
"os/exec"
"strconv"
"strings"
"time"
"github.com/henrygd/beszel/internal/entities/system"
)
const (
// powermetricsSampleIntervalMs is the sampling interval passed to powermetrics (-i).
powermetricsSampleIntervalMs = 500
// powermetricsPollInterval is how often we run powermetrics to collect a new sample.
powermetricsPollInterval = 2 * time.Second
// macmonIntervalMs is the sampling interval passed to macmon pipe (-i), in milliseconds.
macmonIntervalMs = 2500
)
const appleGPUID = "0"
// startPowermetricsCollector runs powermetrics --samplers gpu_power in a loop and updates
// GPU usage and power. Requires root (sudo) on macOS. A single logical GPU is reported as id "0".
func (gm *GPUManager) startPowermetricsCollector() {
// Ensure single GPU entry for Apple GPU
if _, ok := gm.GpuDataMap[appleGPUID]; !ok {
gm.GpuDataMap[appleGPUID] = &system.GPUData{Name: "Apple GPU"}
}
go func() {
failures := 0
for {
if err := gm.collectPowermetrics(); err != nil {
failures++
if failures > maxFailureRetries {
slog.Warn("powermetrics GPU collector failed repeatedly, stopping", "err", err)
break
}
slog.Warn("Error collecting macOS GPU data via powermetrics (may require sudo)", "err", err)
time.Sleep(retryWaitTime)
continue
}
failures = 0
time.Sleep(powermetricsPollInterval)
}
}()
}
// collectPowermetrics runs powermetrics once and parses GPU usage and power from its output.
func (gm *GPUManager) collectPowermetrics() error {
interval := strconv.Itoa(powermetricsSampleIntervalMs)
cmd := exec.Command(powermetricsCmd, "--samplers", "gpu_power", "-i", interval, "-n", "1")
cmd.Stderr = nil
out, err := cmd.Output()
if err != nil {
return err
}
if !gm.parsePowermetricsData(out) {
return errNoValidData
}
return nil
}
// parsePowermetricsData parses powermetrics gpu_power output and updates GpuDataMap["0"].
// Example output:
//
// **** GPU usage ****
// GPU HW active frequency: 444 MHz
// GPU HW active residency: 0.97% (444 MHz: .97% ...
// GPU idle residency: 99.03%
// GPU Power: 4 mW
func (gm *GPUManager) parsePowermetricsData(output []byte) bool {
var idleResidency, powerMW float64
var gotIdle, gotPower bool
scanner := bufio.NewScanner(bytes.NewReader(output))
for scanner.Scan() {
line := strings.TrimSpace(scanner.Text())
if strings.HasPrefix(line, "GPU idle residency:") {
// "GPU idle residency: 99.03%"
fields := strings.Fields(strings.TrimPrefix(line, "GPU idle residency:"))
if len(fields) >= 1 {
pct := strings.TrimSuffix(fields[0], "%")
if v, err := strconv.ParseFloat(pct, 64); err == nil {
idleResidency = v
gotIdle = true
}
}
} else if strings.HasPrefix(line, "GPU Power:") {
// "GPU Power: 4 mW"
fields := strings.Fields(strings.TrimPrefix(line, "GPU Power:"))
if len(fields) >= 1 {
if v, err := strconv.ParseFloat(fields[0], 64); err == nil {
powerMW = v
gotPower = true
}
}
}
}
if err := scanner.Err(); err != nil {
return false
}
if !gotIdle && !gotPower {
return false
}
gm.Lock()
defer gm.Unlock()
if _, ok := gm.GpuDataMap[appleGPUID]; !ok {
gm.GpuDataMap[appleGPUID] = &system.GPUData{Name: "Apple GPU"}
}
gpu := gm.GpuDataMap[appleGPUID]
if gotIdle {
// Usage = 100 - idle residency (e.g. 100 - 99.03 = 0.97%)
gpu.Usage += 100 - idleResidency
}
if gotPower {
// mW -> W
gpu.Power += powerMW / milliwattsInAWatt
}
gpu.Count++
return true
}
// startMacmonCollector runs `macmon pipe` in a loop and parses one JSON object per line.
// This collector does not require sudo. A single logical GPU is reported as id "0".
func (gm *GPUManager) startMacmonCollector() {
if _, ok := gm.GpuDataMap[appleGPUID]; !ok {
gm.GpuDataMap[appleGPUID] = &system.GPUData{Name: "Apple GPU"}
}
go func() {
failures := 0
for {
if err := gm.collectMacmonPipe(); err != nil {
failures++
if failures > maxFailureRetries {
slog.Warn("macmon GPU collector failed repeatedly, stopping", "err", err)
break
}
slog.Warn("Error collecting macOS GPU data via macmon", "err", err)
time.Sleep(retryWaitTime)
continue
}
failures = 0
// `macmon pipe` is long-running; if it returns, wait a bit before restarting.
time.Sleep(retryWaitTime)
}
}()
}
type macmonTemp struct {
GPUTempAvg float64 `json:"gpu_temp_avg"`
}
type macmonSample struct {
GPUPower float64 `json:"gpu_power"` // watts (macmon reports fractional values)
GPURAMPower float64 `json:"gpu_ram_power"` // watts
GPUUsage []float64 `json:"gpu_usage"` // [freq_mhz, usage] where usage is typically 0..1
Temp macmonTemp `json:"temp"`
}
func (gm *GPUManager) collectMacmonPipe() (err error) {
cmd := exec.Command(macmonCmd, "pipe", "-i", strconv.Itoa(macmonIntervalMs))
// Avoid blocking if macmon writes to stderr.
cmd.Stderr = io.Discard
stdout, err := cmd.StdoutPipe()
if err != nil {
return err
}
if err := cmd.Start(); err != nil {
return err
}
// Ensure we always reap the child to avoid zombies on any return path and
// propagate a non-zero exit code if no other error was set.
defer func() {
_ = stdout.Close()
if cmd.ProcessState == nil || !cmd.ProcessState.Exited() {
_ = cmd.Process.Kill()
}
if waitErr := cmd.Wait(); err == nil && waitErr != nil {
err = waitErr
}
}()
scanner := bufio.NewScanner(stdout)
var hadSample bool
for scanner.Scan() {
line := bytes.TrimSpace(scanner.Bytes())
if len(line) == 0 {
continue
}
if gm.parseMacmonLine(line) {
hadSample = true
}
}
if scanErr := scanner.Err(); scanErr != nil {
return scanErr
}
if !hadSample {
return errNoValidData
}
return nil
}
// parseMacmonLine parses a single macmon JSON line and updates Apple GPU metrics.
func (gm *GPUManager) parseMacmonLine(line []byte) bool {
var sample macmonSample
if err := json.Unmarshal(line, &sample); err != nil {
return false
}
usage := 0.0
if len(sample.GPUUsage) >= 2 {
usage = sample.GPUUsage[1]
// Heuristic: macmon typically reports 0..1; convert to percentage.
if usage <= 1.0 {
usage *= 100
}
}
// Consider the line valid if it contains at least one GPU metric.
if usage == 0 && sample.GPUPower == 0 && sample.Temp.GPUTempAvg == 0 {
return false
}
gm.Lock()
defer gm.Unlock()
gpu, ok := gm.GpuDataMap[appleGPUID]
if !ok {
gpu = &system.GPUData{Name: "Apple GPU"}
gm.GpuDataMap[appleGPUID] = gpu
}
gpu.Temperature = sample.Temp.GPUTempAvg
gpu.Usage += usage
// macmon reports power in watts; include VRAM power if present.
gpu.Power += sample.GPUPower + sample.GPURAMPower
gpu.Count++
return true
}

View File

@@ -1,81 +0,0 @@
//go:build darwin
package agent
import (
"testing"
"github.com/henrygd/beszel/internal/entities/system"
"github.com/stretchr/testify/assert"
"github.com/stretchr/testify/require"
)
func TestParsePowermetricsData(t *testing.T) {
input := `
Machine model: Mac14,10
OS version: 25D125
*** Sampled system activity (Sat Feb 14 00:42:06 2026 -0500) (503.05ms elapsed) ***
**** GPU usage ****
GPU HW active frequency: 444 MHz
GPU HW active residency: 0.97% (444 MHz: .97% 612 MHz: 0% 808 MHz: 0% 968 MHz: 0% 1110 MHz: 0% 1236 MHz: 0% 1338 MHz: 0% 1398 MHz: 0%)
GPU SW requested state: (P1 : 100% P2 : 0% P3 : 0% P4 : 0% P5 : 0% P6 : 0% P7 : 0% P8 : 0%)
GPU idle residency: 99.03%
GPU Power: 4 mW
`
gm := &GPUManager{
GpuDataMap: make(map[string]*system.GPUData),
}
valid := gm.parsePowermetricsData([]byte(input))
require.True(t, valid)
g0, ok := gm.GpuDataMap["0"]
require.True(t, ok)
assert.Equal(t, "Apple GPU", g0.Name)
// Usage = 100 - 99.03 = 0.97
assert.InDelta(t, 0.97, g0.Usage, 0.01)
// 4 mW -> 0.004 W
assert.InDelta(t, 0.004, g0.Power, 0.0001)
assert.Equal(t, 1.0, g0.Count)
}
func TestParsePowermetricsDataPartial(t *testing.T) {
// Only power line (e.g. older macOS or different sampler output)
input := `
**** GPU usage ****
GPU Power: 120 mW
`
gm := &GPUManager{
GpuDataMap: make(map[string]*system.GPUData),
}
valid := gm.parsePowermetricsData([]byte(input))
require.True(t, valid)
g0, ok := gm.GpuDataMap["0"]
require.True(t, ok)
assert.Equal(t, "Apple GPU", g0.Name)
assert.InDelta(t, 0.12, g0.Power, 0.001)
assert.Equal(t, 1.0, g0.Count)
}
func TestParseMacmonLine(t *testing.T) {
input := `{"all_power":0.6468324661254883,"ane_power":0.0,"cpu_power":0.6359732151031494,"ecpu_usage":[2061,0.1726151406764984],"gpu_power":0.010859241709113121,"gpu_ram_power":0.000965250947047025,"gpu_usage":[503,0.013633215799927711],"memory":{"ram_total":17179869184,"ram_usage":12322914304,"swap_total":0,"swap_usage":0},"pcpu_usage":[1248,0.11792058497667313],"ram_power":0.14885640144348145,"sys_power":10.4955415725708,"temp":{"cpu_temp_avg":23.041261672973633,"gpu_temp_avg":29.44516944885254},"timestamp":"2026-02-17T19:34:27.942556+00:00"}`
gm := &GPUManager{
GpuDataMap: make(map[string]*system.GPUData),
}
valid := gm.parseMacmonLine([]byte(input))
require.True(t, valid)
g0, ok := gm.GpuDataMap["0"]
require.True(t, ok)
assert.Equal(t, "Apple GPU", g0.Name)
// macmon reports usage fraction 0..1; expect percent conversion.
assert.InDelta(t, 1.3633, g0.Usage, 0.05)
// power includes gpu_power + gpu_ram_power
assert.InDelta(t, 0.011824, g0.Power, 0.0005)
assert.InDelta(t, 29.445, g0.Temperature, 0.01)
assert.Equal(t, 1.0, g0.Count)
}

View File

@@ -1,9 +0,0 @@
//go:build !darwin
package agent
// startPowermetricsCollector is a no-op on non-darwin platforms; the real implementation is in gpu_darwin.go.
func (gm *GPUManager) startPowermetricsCollector() {}
// startMacmonCollector is a no-op on non-darwin platforms; the real implementation is in gpu_darwin.go.
func (gm *GPUManager) startMacmonCollector() {}

View File

@@ -27,11 +27,10 @@ func (gm *GPUManager) updateIntelFromStats(sample *intelGpuStats) bool {
defer gm.Unlock()
// only one gpu for now - cmd doesn't provide all by default
id := "i0" // prefix with i to avoid conflicts with nvidia card ids
gpuData, ok := gm.GpuDataMap[id]
gpuData, ok := gm.GpuDataMap["0"]
if !ok {
gpuData = &system.GPUData{Name: "GPU", Engines: make(map[string]float64)}
gm.GpuDataMap[id] = gpuData
gm.GpuDataMap["0"] = gpuData
}
gpuData.Power += sample.PowerGPU

View File

@@ -1,224 +0,0 @@
//go:build amd64 && (windows || (linux && glibc))
package agent
import (
"fmt"
"log/slog"
"strings"
"time"
"unsafe"
"github.com/ebitengine/purego"
"github.com/henrygd/beszel/internal/entities/system"
)
// NVML constants and types
const (
nvmlSuccess int = 0
)
type nvmlDevice uintptr
type nvmlReturn int
type nvmlMemoryV1 struct {
Total uint64
Free uint64
Used uint64
}
type nvmlMemoryV2 struct {
Version uint32
Total uint64
Reserved uint64
Free uint64
Used uint64
}
type nvmlUtilization struct {
Gpu uint32
Memory uint32
}
type nvmlPciInfo struct {
BusId [16]byte
Domain uint32
Bus uint32
Device uint32
PciDeviceId uint32
PciSubSystemId uint32
}
// NVML function signatures
var (
nvmlInit func() nvmlReturn
nvmlShutdown func() nvmlReturn
nvmlDeviceGetCount func(count *uint32) nvmlReturn
nvmlDeviceGetHandleByIndex func(index uint32, device *nvmlDevice) nvmlReturn
nvmlDeviceGetName func(device nvmlDevice, name *byte, length uint32) nvmlReturn
nvmlDeviceGetMemoryInfo func(device nvmlDevice, memory uintptr) nvmlReturn
nvmlDeviceGetUtilizationRates func(device nvmlDevice, utilization *nvmlUtilization) nvmlReturn
nvmlDeviceGetTemperature func(device nvmlDevice, sensorType int, temp *uint32) nvmlReturn
nvmlDeviceGetPowerUsage func(device nvmlDevice, power *uint32) nvmlReturn
nvmlDeviceGetPciInfo func(device nvmlDevice, pci *nvmlPciInfo) nvmlReturn
nvmlErrorString func(result nvmlReturn) string
)
type nvmlCollector struct {
gm *GPUManager
lib uintptr
devices []nvmlDevice
bdfs []string
isV2 bool
}
func (c *nvmlCollector) init() error {
slog.Debug("NVML: Initializing")
libPath := getNVMLPath()
lib, err := openLibrary(libPath)
if err != nil {
return fmt.Errorf("failed to load %s: %w", libPath, err)
}
c.lib = lib
purego.RegisterLibFunc(&nvmlInit, lib, "nvmlInit")
purego.RegisterLibFunc(&nvmlShutdown, lib, "nvmlShutdown")
purego.RegisterLibFunc(&nvmlDeviceGetCount, lib, "nvmlDeviceGetCount")
purego.RegisterLibFunc(&nvmlDeviceGetHandleByIndex, lib, "nvmlDeviceGetHandleByIndex")
purego.RegisterLibFunc(&nvmlDeviceGetName, lib, "nvmlDeviceGetName")
// Try to get v2 memory info, fallback to v1 if not available
if hasSymbol(lib, "nvmlDeviceGetMemoryInfo_v2") {
c.isV2 = true
purego.RegisterLibFunc(&nvmlDeviceGetMemoryInfo, lib, "nvmlDeviceGetMemoryInfo_v2")
} else {
purego.RegisterLibFunc(&nvmlDeviceGetMemoryInfo, lib, "nvmlDeviceGetMemoryInfo")
}
purego.RegisterLibFunc(&nvmlDeviceGetUtilizationRates, lib, "nvmlDeviceGetUtilizationRates")
purego.RegisterLibFunc(&nvmlDeviceGetTemperature, lib, "nvmlDeviceGetTemperature")
purego.RegisterLibFunc(&nvmlDeviceGetPowerUsage, lib, "nvmlDeviceGetPowerUsage")
purego.RegisterLibFunc(&nvmlDeviceGetPciInfo, lib, "nvmlDeviceGetPciInfo")
purego.RegisterLibFunc(&nvmlErrorString, lib, "nvmlErrorString")
if ret := nvmlInit(); ret != nvmlReturn(nvmlSuccess) {
return fmt.Errorf("nvmlInit failed: %v", ret)
}
var count uint32
if ret := nvmlDeviceGetCount(&count); ret != nvmlReturn(nvmlSuccess) {
return fmt.Errorf("nvmlDeviceGetCount failed: %v", ret)
}
for i := uint32(0); i < count; i++ {
var device nvmlDevice
if ret := nvmlDeviceGetHandleByIndex(i, &device); ret == nvmlReturn(nvmlSuccess) {
c.devices = append(c.devices, device)
// Get BDF for power state check
var pci nvmlPciInfo
if ret := nvmlDeviceGetPciInfo(device, &pci); ret == nvmlReturn(nvmlSuccess) {
busID := string(pci.BusId[:])
if idx := strings.Index(busID, "\x00"); idx != -1 {
busID = busID[:idx]
}
c.bdfs = append(c.bdfs, strings.ToLower(busID))
} else {
c.bdfs = append(c.bdfs, "")
}
}
}
return nil
}
func (c *nvmlCollector) start() {
defer nvmlShutdown()
ticker := time.Tick(3 * time.Second)
for range ticker {
c.collect()
}
}
func (c *nvmlCollector) collect() {
c.gm.Lock()
defer c.gm.Unlock()
for i, device := range c.devices {
id := fmt.Sprintf("%d", i)
bdf := c.bdfs[i]
// Update GPUDataMap
if _, ok := c.gm.GpuDataMap[id]; !ok {
var nameBuf [64]byte
if ret := nvmlDeviceGetName(device, &nameBuf[0], 64); ret != nvmlReturn(nvmlSuccess) {
continue
}
name := string(nameBuf[:strings.Index(string(nameBuf[:]), "\x00")])
name = strings.TrimPrefix(name, "NVIDIA ")
c.gm.GpuDataMap[id] = &system.GPUData{Name: strings.TrimSuffix(name, " Laptop GPU")}
}
gpu := c.gm.GpuDataMap[id]
if bdf != "" && !c.isGPUActive(bdf) {
slog.Debug("NVML: GPU is suspended, skipping", "bdf", bdf)
gpu.Temperature = 0
gpu.MemoryUsed = 0
continue
}
// Utilization
var utilization nvmlUtilization
if ret := nvmlDeviceGetUtilizationRates(device, &utilization); ret != nvmlReturn(nvmlSuccess) {
slog.Debug("NVML: Utilization failed (GPU likely suspended)", "bdf", bdf, "ret", ret)
gpu.Temperature = 0
gpu.MemoryUsed = 0
continue
}
slog.Debug("NVML: Collecting data for GPU", "bdf", bdf)
// Temperature
var temp uint32
nvmlDeviceGetTemperature(device, 0, &temp) // 0 is NVML_TEMPERATURE_GPU
// Memory: only poll if GPU is active to avoid leaving D3cold state (#1522)
if utilization.Gpu > 0 {
var usedMem, totalMem uint64
if c.isV2 {
var memory nvmlMemoryV2
memory.Version = 0x02000028 // (2 << 24) | 40 bytes
if ret := nvmlDeviceGetMemoryInfo(device, uintptr(unsafe.Pointer(&memory))); ret != nvmlReturn(nvmlSuccess) {
slog.Debug("NVML: MemoryInfo_v2 failed", "bdf", bdf, "ret", ret)
} else {
usedMem = memory.Used
totalMem = memory.Total
}
} else {
var memory nvmlMemoryV1
if ret := nvmlDeviceGetMemoryInfo(device, uintptr(unsafe.Pointer(&memory))); ret != nvmlReturn(nvmlSuccess) {
slog.Debug("NVML: MemoryInfo failed", "bdf", bdf, "ret", ret)
} else {
usedMem = memory.Used
totalMem = memory.Total
}
}
if totalMem > 0 {
gpu.MemoryUsed = float64(usedMem) / 1024 / 1024 / mebibytesInAMegabyte
gpu.MemoryTotal = float64(totalMem) / 1024 / 1024 / mebibytesInAMegabyte
}
} else {
slog.Debug("NVML: Skipping memory info (utilization=0)", "bdf", bdf)
}
// Power
var power uint32
nvmlDeviceGetPowerUsage(device, &power)
gpu.Temperature = float64(temp)
gpu.Usage += float64(utilization.Gpu)
gpu.Power += float64(power) / 1000.0
gpu.Count++
slog.Debug("NVML: Collected data", "gpu", gpu)
}
}

View File

@@ -1,57 +0,0 @@
//go:build glibc && linux && amd64
package agent
import (
"log/slog"
"os"
"path/filepath"
"strings"
"github.com/ebitengine/purego"
)
func openLibrary(name string) (uintptr, error) {
return purego.Dlopen(name, purego.RTLD_NOW|purego.RTLD_GLOBAL)
}
func getNVMLPath() string {
return "libnvidia-ml.so.1"
}
func hasSymbol(lib uintptr, symbol string) bool {
_, err := purego.Dlsym(lib, symbol)
return err == nil
}
func (c *nvmlCollector) isGPUActive(bdf string) bool {
// runtime_status
statusPath := filepath.Join("/sys/bus/pci/devices", bdf, "power/runtime_status")
status, err := os.ReadFile(statusPath)
if err != nil {
slog.Debug("NVML: Can't read runtime_status", "bdf", bdf, "err", err)
return true // Assume active if we can't read status
}
statusStr := strings.TrimSpace(string(status))
if statusStr != "active" && statusStr != "resuming" {
slog.Debug("NVML: GPU not active", "bdf", bdf, "status", statusStr)
return false
}
// power_state (D0 check)
// Find any drm card device power_state
pstatePathPattern := filepath.Join("/sys/bus/pci/devices", bdf, "drm/card*/device/power_state")
matches, _ := filepath.Glob(pstatePathPattern)
if len(matches) > 0 {
pstate, err := os.ReadFile(matches[0])
if err == nil {
pstateStr := strings.TrimSpace(string(pstate))
if pstateStr != "D0" {
slog.Debug("NVML: GPU not in D0 state", "bdf", bdf, "pstate", pstateStr)
return false
}
}
}
return true
}

View File

@@ -1,15 +0,0 @@
//go:build (!linux && !windows) || !amd64 || (linux && !glibc)
package agent
import "fmt"
type nvmlCollector struct {
gm *GPUManager
}
func (c *nvmlCollector) init() error {
return fmt.Errorf("nvml not supported on this platform")
}
func (c *nvmlCollector) start() {}

View File

@@ -1,25 +0,0 @@
//go:build windows && amd64
package agent
import (
"golang.org/x/sys/windows"
)
func openLibrary(name string) (uintptr, error) {
handle, err := windows.LoadLibrary(name)
return uintptr(handle), err
}
func getNVMLPath() string {
return "nvml.dll"
}
func hasSymbol(lib uintptr, symbol string) bool {
_, err := windows.GetProcAddress(windows.Handle(lib), symbol)
return err == nil
}
func (c *nvmlCollector) isGPUActive(bdf string) bool {
return true
}

View File

@@ -1,159 +0,0 @@
package agent
import (
"encoding/json"
"io"
"log/slog"
"os/exec"
"strconv"
"strings"
"time"
"github.com/henrygd/beszel/internal/entities/system"
)
type nvtopSnapshot struct {
DeviceName string `json:"device_name"`
Temp *string `json:"temp"`
PowerDraw *string `json:"power_draw"`
GpuUtil *string `json:"gpu_util"`
MemTotal *string `json:"mem_total"`
MemUsed *string `json:"mem_used"`
}
// parseNvtopNumber parses nvtop numeric strings with units (C/W/%).
func parseNvtopNumber(raw string) float64 {
cleaned := strings.TrimSpace(raw)
cleaned = strings.TrimSuffix(cleaned, "C")
cleaned = strings.TrimSuffix(cleaned, "W")
cleaned = strings.TrimSuffix(cleaned, "%")
val, _ := strconv.ParseFloat(cleaned, 64)
return val
}
// parseNvtopData parses a single nvtop JSON snapshot payload.
func (gm *GPUManager) parseNvtopData(output []byte) bool {
var snapshots []nvtopSnapshot
if err := json.Unmarshal(output, &snapshots); err != nil || len(snapshots) == 0 {
return false
}
return gm.updateNvtopSnapshots(snapshots)
}
// updateNvtopSnapshots applies one decoded nvtop snapshot batch to GPU accumulators.
func (gm *GPUManager) updateNvtopSnapshots(snapshots []nvtopSnapshot) bool {
gm.Lock()
defer gm.Unlock()
valid := false
usedIDs := make(map[string]struct{}, len(snapshots))
for i, sample := range snapshots {
if sample.DeviceName == "" {
continue
}
indexID := "n" + strconv.Itoa(i)
id := indexID
// nvtop ordering can change, so prefer reusing an existing slot with matching device name.
if existingByIndex, ok := gm.GpuDataMap[indexID]; ok && existingByIndex.Name != "" && existingByIndex.Name != sample.DeviceName {
for existingID, gpu := range gm.GpuDataMap {
if !strings.HasPrefix(existingID, "n") {
continue
}
if _, taken := usedIDs[existingID]; taken {
continue
}
if gpu.Name == sample.DeviceName {
id = existingID
break
}
}
}
if _, ok := gm.GpuDataMap[id]; !ok {
gm.GpuDataMap[id] = &system.GPUData{Name: sample.DeviceName}
}
gpu := gm.GpuDataMap[id]
gpu.Name = sample.DeviceName
if sample.Temp != nil {
gpu.Temperature = parseNvtopNumber(*sample.Temp)
}
if sample.MemUsed != nil {
gpu.MemoryUsed = bytesToMegabytes(parseNvtopNumber(*sample.MemUsed))
}
if sample.MemTotal != nil {
gpu.MemoryTotal = bytesToMegabytes(parseNvtopNumber(*sample.MemTotal))
}
if sample.GpuUtil != nil {
gpu.Usage += parseNvtopNumber(*sample.GpuUtil)
}
if sample.PowerDraw != nil {
gpu.Power += parseNvtopNumber(*sample.PowerDraw)
}
gpu.Count++
usedIDs[id] = struct{}{}
valid = true
}
return valid
}
// collectNvtopStats runs nvtop loop mode and continuously decodes JSON snapshots.
func (gm *GPUManager) collectNvtopStats(interval string) error {
cmd := exec.Command(nvtopCmd, "-lP", "-d", interval)
stdout, err := cmd.StdoutPipe()
if err != nil {
return err
}
if err := cmd.Start(); err != nil {
return err
}
defer func() {
_ = stdout.Close()
if cmd.ProcessState == nil || !cmd.ProcessState.Exited() {
_ = cmd.Process.Kill()
}
_ = cmd.Wait()
}()
decoder := json.NewDecoder(stdout)
foundValid := false
for {
var snapshots []nvtopSnapshot
if err := decoder.Decode(&snapshots); err != nil {
if err == io.EOF {
if foundValid {
return nil
}
return errNoValidData
}
return err
}
if gm.updateNvtopSnapshots(snapshots) {
foundValid = true
}
}
}
// startNvtopCollector starts nvtop collection with retry or fallback callback handling.
func (gm *GPUManager) startNvtopCollector(interval string, onFailure func()) {
go func() {
failures := 0
for {
if err := gm.collectNvtopStats(interval); err != nil {
if onFailure != nil {
slog.Warn("Error collecting GPU data via nvtop", "err", err)
onFailure()
return
}
failures++
if failures > maxFailureRetries {
break
}
slog.Warn("Error collecting GPU data via nvtop", "err", err)
time.Sleep(retryWaitTime)
continue
}
}
}()
}

View File

@@ -1,4 +1,5 @@
//go:build testing
// +build testing
package agent
@@ -249,100 +250,6 @@ func TestParseAmdData(t *testing.T) {
}
}
func TestParseNvtopData(t *testing.T) {
input, err := os.ReadFile("test-data/nvtop.json")
require.NoError(t, err)
gm := &GPUManager{
GpuDataMap: make(map[string]*system.GPUData),
}
valid := gm.parseNvtopData(input)
require.True(t, valid)
g0, ok := gm.GpuDataMap["n0"]
require.True(t, ok)
assert.Equal(t, "NVIDIA GeForce RTX 3050 Ti Laptop GPU", g0.Name)
assert.Equal(t, 48.0, g0.Temperature)
assert.Equal(t, 5.0, g0.Usage)
assert.Equal(t, 13.0, g0.Power)
assert.Equal(t, bytesToMegabytes(349372416), g0.MemoryUsed)
assert.Equal(t, bytesToMegabytes(4294967296), g0.MemoryTotal)
assert.Equal(t, 1.0, g0.Count)
g1, ok := gm.GpuDataMap["n1"]
require.True(t, ok)
assert.Equal(t, "AMD Radeon 680M", g1.Name)
assert.Equal(t, 48.0, g1.Temperature)
assert.Equal(t, 12.0, g1.Usage)
assert.Equal(t, 9.0, g1.Power)
assert.Equal(t, bytesToMegabytes(1213784064), g1.MemoryUsed)
assert.Equal(t, bytesToMegabytes(16929173504), g1.MemoryTotal)
assert.Equal(t, 1.0, g1.Count)
}
func TestUpdateNvtopSnapshotsKeepsDeviceAssociationWhenOrderChanges(t *testing.T) {
strPtr := func(s string) *string { return &s }
gm := &GPUManager{
GpuDataMap: make(map[string]*system.GPUData),
}
firstBatch := []nvtopSnapshot{
{
DeviceName: "NVIDIA GeForce RTX 3050 Ti Laptop GPU",
GpuUtil: strPtr("20%"),
PowerDraw: strPtr("10W"),
},
{
DeviceName: "AMD Radeon 680M",
GpuUtil: strPtr("30%"),
PowerDraw: strPtr("20W"),
},
}
secondBatchSwapped := []nvtopSnapshot{
{
DeviceName: "AMD Radeon 680M",
GpuUtil: strPtr("40%"),
PowerDraw: strPtr("25W"),
},
{
DeviceName: "NVIDIA GeForce RTX 3050 Ti Laptop GPU",
GpuUtil: strPtr("50%"),
PowerDraw: strPtr("15W"),
},
}
require.True(t, gm.updateNvtopSnapshots(firstBatch))
require.True(t, gm.updateNvtopSnapshots(secondBatchSwapped))
nvidia := gm.GpuDataMap["n0"]
require.NotNil(t, nvidia)
assert.Equal(t, "NVIDIA GeForce RTX 3050 Ti Laptop GPU", nvidia.Name)
assert.Equal(t, 70.0, nvidia.Usage)
assert.Equal(t, 25.0, nvidia.Power)
assert.Equal(t, 2.0, nvidia.Count)
amd := gm.GpuDataMap["n1"]
require.NotNil(t, amd)
assert.Equal(t, "AMD Radeon 680M", amd.Name)
assert.Equal(t, 70.0, amd.Usage)
assert.Equal(t, 45.0, amd.Power)
assert.Equal(t, 2.0, amd.Count)
}
func TestParseCollectorPriority(t *testing.T) {
got := parseCollectorPriority(" nvml, nvidia-smi, intel_gpu_top, amd_sysfs, nvtop, rocm-smi, bad ")
want := []collectorSource{
collectorSourceNVML,
collectorSourceNvidiaSMI,
collectorSourceIntelGpuTop,
collectorSourceAmdSysfs,
collectorSourceNVTop,
collectorSourceRocmSMI,
}
assert.Equal(t, want, got)
}
func TestParseJetsonData(t *testing.T) {
tests := []struct {
name string
@@ -400,19 +307,6 @@ func TestParseJetsonData(t *testing.T) {
Count: 1,
},
},
{
name: "orin-style output with GPU@ temp and VDD_SYS_GPU power",
input: "RAM 3276/7859MB (lfb 5x4MB) SWAP 1626/12122MB (cached 181MB) CPU [44%@1421,49%@2031,67%@2034,17%@1420,25%@1419,8%@1420] EMC_FREQ 1%@1866 GR3D_FREQ 0%@114 APE 150 MTS fg 1% bg 1% PLL@42.5C MCPU@42.5C PMIC@50C Tboard@38C GPU@39.5C BCPU@42.5C thermal@41.3C Tdiode@39.25C VDD_SYS_GPU 182/182 VDD_SYS_SOC 730/730 VDD_4V0_WIFI 0/0 VDD_IN 5297/5297 VDD_SYS_CPU 1917/1917 VDD_SYS_DDR 1241/1241",
wantMetrics: &system.GPUData{
Name: "GPU",
MemoryUsed: 3276.0,
MemoryTotal: 7859.0,
Usage: 0.0,
Power: 0.182, // 182mW -> 0.182W
Temperature: 39.5,
Count: 1,
},
},
}
for _, tt := range tests {
@@ -931,7 +825,7 @@ func TestInitializeSnapshots(t *testing.T) {
}
func TestCalculateGPUAverage(t *testing.T) {
t.Run("returns cached average when deltaCount is zero", func(t *testing.T) {
t.Run("returns zero value when deltaCount is zero", func(t *testing.T) {
gm := &GPUManager{
lastSnapshots: map[uint16]map[string]*gpuSnapshot{
5000: {
@@ -944,10 +838,9 @@ func TestCalculateGPUAverage(t *testing.T) {
}
gpu := &system.GPUData{
Count: 10.0, // Same as snapshot, so delta = 0
Usage: 100.0,
Power: 200.0,
Temperature: 50.0, // Non-zero to avoid "suspended" check
Count: 10.0, // Same as snapshot, so delta = 0
Usage: 100.0,
Power: 200.0,
}
result := gm.calculateGPUAverage("0", gpu, 5000)
@@ -956,31 +849,6 @@ func TestCalculateGPUAverage(t *testing.T) {
assert.Equal(t, 100.0, result.Power, "Should return cached average")
})
t.Run("returns zero value when GPU is suspended", func(t *testing.T) {
gm := &GPUManager{
lastSnapshots: map[uint16]map[string]*gpuSnapshot{
5000: {
"0": {count: 10, usage: 100, power: 200},
},
},
lastAvgData: map[string]system.GPUData{
"0": {Usage: 50.0, Power: 100.0},
},
}
gpu := &system.GPUData{
Name: "Test GPU",
Count: 10.0,
Temperature: 0,
MemoryUsed: 0,
}
result := gm.calculateGPUAverage("0", gpu, 5000)
assert.Equal(t, 0.0, result.Usage, "Should return zero usage")
assert.Equal(t, 0.0, result.Power, "Should return zero power")
})
t.Run("calculates average for standard GPU", func(t *testing.T) {
gm := &GPUManager{
lastSnapshots: map[uint16]map[string]*gpuSnapshot{
@@ -1080,35 +948,36 @@ func TestCalculateGPUAverage(t *testing.T) {
})
}
func TestGPUCapabilitiesAndLegacyPriority(t *testing.T) {
func TestDetectGPUs(t *testing.T) {
// Save original PATH
origPath := os.Getenv("PATH")
defer os.Setenv("PATH", origPath)
hasAmdSysfs := (&GPUManager{}).hasAmdSysfs()
// Set up temp dir with the commands
tempDir := t.TempDir()
os.Setenv("PATH", tempDir)
tests := []struct {
name string
setupCommands func(string) error
setupCommands func() error
wantNvidiaSmi bool
wantRocmSmi bool
wantTegrastats bool
wantNvtop bool
wantErr bool
}{
{
name: "nvidia-smi not available",
setupCommands: func(_ string) error {
setupCommands: func() error {
return nil
},
wantNvidiaSmi: false,
wantRocmSmi: false,
wantTegrastats: false,
wantNvtop: false,
wantErr: true,
},
{
name: "nvidia-smi available",
setupCommands: func(tempDir string) error {
setupCommands: func() error {
path := filepath.Join(tempDir, "nvidia-smi")
script := `#!/bin/sh
echo "test"`
@@ -1120,12 +989,11 @@ echo "test"`
wantNvidiaSmi: true,
wantTegrastats: false,
wantRocmSmi: false,
wantNvtop: false,
wantErr: false,
},
{
name: "rocm-smi available",
setupCommands: func(tempDir string) error {
setupCommands: func() error {
path := filepath.Join(tempDir, "rocm-smi")
script := `#!/bin/sh
echo "test"`
@@ -1134,15 +1002,14 @@ echo "test"`
}
return nil
},
wantNvidiaSmi: false,
wantNvidiaSmi: true,
wantRocmSmi: true,
wantTegrastats: false,
wantNvtop: false,
wantErr: false,
},
{
name: "tegrastats available",
setupCommands: func(tempDir string) error {
setupCommands: func() error {
path := filepath.Join(tempDir, "tegrastats")
script := `#!/bin/sh
echo "test"`
@@ -1152,31 +1019,13 @@ echo "test"`
return nil
},
wantNvidiaSmi: false,
wantRocmSmi: false,
wantRocmSmi: true,
wantTegrastats: true,
wantNvtop: false,
wantErr: false,
},
{
name: "nvtop available",
setupCommands: func(tempDir string) error {
path := filepath.Join(tempDir, "nvtop")
script := `#!/bin/sh
echo "[]"`
if err := os.WriteFile(path, []byte(script), 0755); err != nil {
return err
}
return nil
},
wantNvidiaSmi: false,
wantRocmSmi: false,
wantTegrastats: false,
wantNvtop: true,
wantErr: false,
},
{
name: "no gpu tools available",
setupCommands: func(_ string) error {
setupCommands: func() error {
os.Setenv("PATH", "")
return nil
},
@@ -1186,53 +1035,29 @@ echo "[]"`
for _, tt := range tests {
t.Run(tt.name, func(t *testing.T) {
tempDir := t.TempDir()
os.Setenv("PATH", tempDir)
if err := tt.setupCommands(tempDir); err != nil {
if err := tt.setupCommands(); err != nil {
t.Fatal(err)
}
gm := &GPUManager{}
caps := gm.discoverGpuCapabilities()
var err error
if !hasAnyGpuCollector(caps) {
err = fmt.Errorf(noGPUFoundMsg)
}
priorities := gm.resolveLegacyCollectorPriority(caps)
hasPriority := func(source collectorSource) bool {
for _, s := range priorities {
if s == source {
return true
}
}
return false
}
gotNvidiaSmi := hasPriority(collectorSourceNvidiaSMI)
gotRocmSmi := hasPriority(collectorSourceRocmSMI)
gotTegrastats := caps.hasTegrastats
gotNvtop := caps.hasNvtop
err := gm.detectGPUs()
t.Logf("nvidiaSmi: %v, rocmSmi: %v, tegrastats: %v", gotNvidiaSmi, gotRocmSmi, gotTegrastats)
t.Logf("nvidiaSmi: %v, rocmSmi: %v, tegrastats: %v", gm.nvidiaSmi, gm.rocmSmi, gm.tegrastats)
wantErr := tt.wantErr
if hasAmdSysfs && (tt.name == "nvidia-smi not available" || tt.name == "no gpu tools available") {
wantErr = false
}
if wantErr {
if tt.wantErr {
assert.Error(t, err)
return
}
assert.NoError(t, err)
assert.Equal(t, tt.wantNvidiaSmi, gotNvidiaSmi)
assert.Equal(t, tt.wantRocmSmi, gotRocmSmi)
assert.Equal(t, tt.wantTegrastats, gotTegrastats)
assert.Equal(t, tt.wantNvtop, gotNvtop)
assert.Equal(t, tt.wantNvidiaSmi, gm.nvidiaSmi)
assert.Equal(t, tt.wantRocmSmi, gm.rocmSmi)
assert.Equal(t, tt.wantTegrastats, gm.tegrastats)
})
}
}
func TestCollectorStartHelpers(t *testing.T) {
func TestStartCollector(t *testing.T) {
// Save original PATH
origPath := os.Getenv("PATH")
defer os.Setenv("PATH", origPath)
@@ -1317,27 +1142,6 @@ echo "11-14-2024 22:54:33 RAM 1024/4096MB GR3D_FREQ 80% tj@70C VDD_GPU_SOC 1000m
},
},
},
{
name: "nvtop collector",
command: "nvtop",
setup: func(t *testing.T) error {
path := filepath.Join(dir, "nvtop")
script := `#!/bin/sh
echo '[{"device_name":"NVIDIA Test GPU","temp":"52C","power_draw":"31W","gpu_util":"37%","mem_total":"4294967296","mem_used":"536870912","processes":[]}]'`
if err := os.WriteFile(path, []byte(script), 0755); err != nil {
return err
}
return nil
},
validate: func(t *testing.T, gm *GPUManager) {
gpu, exists := gm.GpuDataMap["n0"]
assert.True(t, exists)
if exists {
assert.Equal(t, "NVIDIA Test GPU", gpu.Name)
assert.Equal(t, 52.0, gpu.Temperature)
}
},
},
}
for _, tt := range tests {
@@ -1350,157 +1154,13 @@ echo '[{"device_name":"NVIDIA Test GPU","temp":"52C","power_draw":"31W","gpu_uti
GpuDataMap: make(map[string]*system.GPUData),
}
}
switch tt.command {
case nvidiaSmiCmd:
tt.gm.startNvidiaSmiCollector("4")
case rocmSmiCmd:
tt.gm.startRocmSmiCollector(4300 * time.Millisecond)
case tegraStatsCmd:
tt.gm.startTegraStatsCollector("3700")
case nvtopCmd:
tt.gm.startNvtopCollector("30", nil)
default:
t.Fatalf("unknown test command %q", tt.command)
}
tt.gm.startCollector(tt.command)
time.Sleep(50 * time.Millisecond) // Give collector time to run
tt.validate(t, tt.gm)
})
}
}
func TestNewGPUManagerPriorityNvtopFallback(t *testing.T) {
origPath := os.Getenv("PATH")
defer os.Setenv("PATH", origPath)
dir := t.TempDir()
os.Setenv("PATH", dir)
t.Setenv("BESZEL_AGENT_GPU_COLLECTOR", "nvtop,nvidia-smi")
nvtopPath := filepath.Join(dir, "nvtop")
nvtopScript := `#!/bin/sh
echo 'not-json'`
require.NoError(t, os.WriteFile(nvtopPath, []byte(nvtopScript), 0755))
nvidiaPath := filepath.Join(dir, "nvidia-smi")
nvidiaScript := `#!/bin/sh
echo "0, NVIDIA Priority GPU, 45, 512, 2048, 12, 25"`
require.NoError(t, os.WriteFile(nvidiaPath, []byte(nvidiaScript), 0755))
gm, err := NewGPUManager()
require.NoError(t, err)
require.NotNil(t, gm)
time.Sleep(150 * time.Millisecond)
gpu, ok := gm.GpuDataMap["0"]
require.True(t, ok)
assert.Equal(t, "Priority GPU", gpu.Name)
assert.Equal(t, 45.0, gpu.Temperature)
}
func TestNewGPUManagerPriorityMixedCollectors(t *testing.T) {
origPath := os.Getenv("PATH")
defer os.Setenv("PATH", origPath)
dir := t.TempDir()
os.Setenv("PATH", dir)
t.Setenv("BESZEL_AGENT_GPU_COLLECTOR", "intel_gpu_top,rocm-smi")
intelPath := filepath.Join(dir, "intel_gpu_top")
intelScript := `#!/bin/sh
echo "Freq MHz IRQ RC6 Power W IMC MiB/s RCS VCS"
echo " req act /s % gpu pkg rd wr % se wa % se wa"
echo "226 223 338 58 2.00 2.69 1820 965 0.00 0 0 0.00 0 0"
echo "189 187 412 67 1.80 2.45 1950 823 8.50 2 1 15.00 1 0"
`
require.NoError(t, os.WriteFile(intelPath, []byte(intelScript), 0755))
rocmPath := filepath.Join(dir, "rocm-smi")
rocmScript := `#!/bin/sh
echo '{"card0": {"Temperature (Sensor edge) (C)": "49.0", "Current Socket Graphics Package Power (W)": "28.159", "GPU use (%)": "0", "VRAM Total Memory (B)": "536870912", "VRAM Total Used Memory (B)": "445550592", "Card Series": "Rembrandt [Radeon 680M]", "GUID": "34756"}}'
`
require.NoError(t, os.WriteFile(rocmPath, []byte(rocmScript), 0755))
gm, err := NewGPUManager()
require.NoError(t, err)
require.NotNil(t, gm)
time.Sleep(150 * time.Millisecond)
_, intelOk := gm.GpuDataMap["i0"]
_, amdOk := gm.GpuDataMap["34756"]
assert.True(t, intelOk)
assert.True(t, amdOk)
}
func TestNewGPUManagerPriorityNvmlFallbackToNvidiaSmi(t *testing.T) {
origPath := os.Getenv("PATH")
defer os.Setenv("PATH", origPath)
dir := t.TempDir()
os.Setenv("PATH", dir)
t.Setenv("BESZEL_AGENT_GPU_COLLECTOR", "nvml,nvidia-smi")
nvidiaPath := filepath.Join(dir, "nvidia-smi")
nvidiaScript := `#!/bin/sh
echo "0, NVIDIA Fallback GPU, 41, 256, 1024, 8, 14"`
require.NoError(t, os.WriteFile(nvidiaPath, []byte(nvidiaScript), 0755))
gm, err := NewGPUManager()
require.NoError(t, err)
require.NotNil(t, gm)
time.Sleep(150 * time.Millisecond)
gpu, ok := gm.GpuDataMap["0"]
require.True(t, ok)
assert.Equal(t, "Fallback GPU", gpu.Name)
}
func TestNewGPUManagerConfiguredCollectorsMustStart(t *testing.T) {
origPath := os.Getenv("PATH")
defer os.Setenv("PATH", origPath)
dir := t.TempDir()
os.Setenv("PATH", dir)
t.Run("configured valid collector unavailable", func(t *testing.T) {
t.Setenv("BESZEL_AGENT_GPU_COLLECTOR", "nvidia-smi")
gm, err := NewGPUManager()
require.Nil(t, gm)
require.Error(t, err)
assert.Contains(t, err.Error(), "no configured GPU collectors are available")
})
t.Run("configured collector list has only unknown entries", func(t *testing.T) {
t.Setenv("BESZEL_AGENT_GPU_COLLECTOR", "bad,unknown")
gm, err := NewGPUManager()
require.Nil(t, gm)
require.Error(t, err)
assert.Contains(t, err.Error(), "no configured GPU collectors are available")
})
}
func TestNewGPUManagerJetsonIgnoresCollectorConfig(t *testing.T) {
origPath := os.Getenv("PATH")
defer os.Setenv("PATH", origPath)
dir := t.TempDir()
os.Setenv("PATH", dir)
t.Setenv("BESZEL_AGENT_GPU_COLLECTOR", "nvidia-smi")
tegraPath := filepath.Join(dir, "tegrastats")
tegraScript := `#!/bin/sh
echo "11-14-2024 22:54:33 RAM 1024/4096MB GR3D_FREQ 80% tj@70C VDD_GPU_SOC 1000mW"`
require.NoError(t, os.WriteFile(tegraPath, []byte(tegraScript), 0755))
gm, err := NewGPUManager()
require.NoError(t, err)
require.NotNil(t, gm)
time.Sleep(100 * time.Millisecond)
gpu, ok := gm.GpuDataMap["0"]
require.True(t, ok)
assert.Equal(t, "GPU", gpu.Name)
}
// TestAccumulationTableDriven tests the accumulation behavior for all three GPU types
func TestAccumulation(t *testing.T) {
type expectedGPUValues struct {
@@ -1686,7 +1346,7 @@ func TestIntelUpdateFromStats(t *testing.T) {
ok := gm.updateIntelFromStats(&sample1)
assert.True(t, ok)
gpu := gm.GpuDataMap["i0"]
gpu := gm.GpuDataMap["0"]
require.NotNil(t, gpu)
assert.Equal(t, "GPU", gpu.Name)
assert.EqualValues(t, 10.5, gpu.Power)
@@ -1708,7 +1368,7 @@ func TestIntelUpdateFromStats(t *testing.T) {
ok = gm.updateIntelFromStats(&sample2)
assert.True(t, ok)
gpu = gm.GpuDataMap["i0"]
gpu = gm.GpuDataMap["0"]
require.NotNil(t, gpu)
assert.EqualValues(t, 10.5, gpu.Power)
assert.EqualValues(t, 30.0, gpu.Engines["Render/3D"]) // 20 + 10
@@ -1747,7 +1407,7 @@ echo "298 295 278 51 2.20 3.12 1675 942 5.75 1 2 9.50
t.Fatalf("collectIntelStats error: %v", err)
}
gpu := gm.GpuDataMap["i0"]
gpu := gm.GpuDataMap["0"]
require.NotNil(t, gpu)
// Power should be sum of samples 2-4 (first is skipped): 2.0 + 1.8 + 2.2 = 6.0
assert.EqualValues(t, 6.0, gpu.Power)

View File

@@ -9,7 +9,7 @@ import (
"github.com/henrygd/beszel/internal/common"
"github.com/henrygd/beszel/internal/entities/smart"
"log/slog"
"golang.org/x/exp/slog"
)
// HandlerContext provides context for request handlers
@@ -94,7 +94,7 @@ func (h *GetDataHandler) Handle(hctx *HandlerContext) error {
var options common.DataRequestOptions
_ = cbor.Unmarshal(hctx.Request.Data, &options)
sysStats := hctx.Agent.gatherStats(options)
sysStats := hctx.Agent.gatherStats(options.CacheTimeMs)
return hctx.SendResponse(sysStats, hctx.RequestID)
}

View File

@@ -1,4 +1,5 @@
//go:build testing
// +build testing
package agent

View File

@@ -9,31 +9,11 @@ import (
"log"
"os"
"path/filepath"
"runtime"
"time"
)
// healthFile is the path to the health file
var healthFile = getHealthFilePath()
func getHealthFilePath() string {
filename := "beszel_health"
if runtime.GOOS == "linux" {
fullPath := filepath.Join("/dev/shm", filename)
if err := updateHealthFile(fullPath); err == nil {
return fullPath
}
}
return filepath.Join(os.TempDir(), filename)
}
func updateHealthFile(path string) error {
file, err := os.Create(path)
if err != nil {
return err
}
return file.Close()
}
var healthFile = filepath.Join(os.TempDir(), "beszel_health")
// Check checks if the agent is connected by checking the modification time of the health file
func Check() error {
@@ -50,7 +30,11 @@ func Check() error {
// Update updates the modification time of the health file
func Update() error {
return updateHealthFile(healthFile)
file, err := os.Create(healthFile)
if err != nil {
return err
}
return file.Close()
}
// CleanUp removes the health file

View File

@@ -1,4 +1,5 @@
//go:build testing
// +build testing
package health
@@ -36,6 +37,7 @@ func TestHealth(t *testing.T) {
})
// This test uses synctest to simulate time passing.
// NOTE: This test requires GOEXPERIMENT=synctest to run.
t.Run("check with simulated time", func(t *testing.T) {
synctest.Test(t, func(t *testing.T) {
// Update the file to set the initial timestamp.

View File

@@ -52,12 +52,7 @@ class Program
foreach (var sensor in hardware.Sensors)
{
var validTemp = sensor.SensorType == SensorType.Temperature && sensor.Value.HasValue;
if (!validTemp ||
sensor.Name.IndexOf("Distance", StringComparison.OrdinalIgnoreCase) >= 0 ||
sensor.Name.IndexOf("Limit", StringComparison.OrdinalIgnoreCase) >= 0 ||
sensor.Name.IndexOf("Critical", StringComparison.OrdinalIgnoreCase) >= 0 ||
sensor.Name.IndexOf("Warning", StringComparison.OrdinalIgnoreCase) >= 0 ||
sensor.Name.IndexOf("Resolution", StringComparison.OrdinalIgnoreCase) >= 0)
if (!validTemp || sensor.Name.Contains("Distance"))
{
continue;
}

View File

@@ -3,11 +3,9 @@
<OutputType>Exe</OutputType>
<TargetFramework>net48</TargetFramework>
<Platforms>x64</Platforms>
<RuntimeIdentifier>win-x64</RuntimeIdentifier>
<AppendRuntimeIdentifierToOutputPath>false</AppendRuntimeIdentifierToOutputPath>
</PropertyGroup>
<ItemGroup>
<PackageReference Include="LibreHardwareMonitorLib" Version="0.9.5" />
<PackageReference Include="LibreHardwareMonitorLib" Version="0.9.4" />
</ItemGroup>
</Project>

View File

@@ -1,31 +0,0 @@
package agent
import (
"github.com/fxamacker/cbor/v2"
"github.com/henrygd/beszel/internal/common"
"github.com/henrygd/beszel/internal/entities/smart"
"github.com/henrygd/beszel/internal/entities/system"
"github.com/henrygd/beszel/internal/entities/systemd"
)
// newAgentResponse creates an AgentResponse using legacy typed fields.
// This maintains backward compatibility with <= 0.17 hubs that expect specific fields.
func newAgentResponse(data any, requestID *uint32) common.AgentResponse {
response := common.AgentResponse{Id: requestID}
switch v := data.(type) {
case *system.CombinedData:
response.SystemData = v
case *common.FingerprintResponse:
response.Fingerprint = v
case string:
response.String = &v
case map[string]smart.SmartData:
response.SmartData = v
case systemd.ServiceDetails:
response.ServiceInfo = v
default:
// For unknown types, use the generic Data field
response.Data, _ = cbor.Marshal(data)
}
return response
}

View File

@@ -1,4 +1,5 @@
//go:build testing
// +build testing
package agent

View File

@@ -13,7 +13,9 @@ import (
"github.com/henrygd/beszel"
"github.com/henrygd/beszel/internal/common"
"github.com/henrygd/beszel/internal/entities/smart"
"github.com/henrygd/beszel/internal/entities/system"
"github.com/henrygd/beszel/internal/entities/systemd"
"github.com/blang/semver"
"github.com/fxamacker/cbor/v2"
@@ -36,9 +38,6 @@ var hubVersions map[string]semver.Version
// and begins listening for connections. Returns an error if the server
// is already running or if there's an issue starting the server.
func (a *Agent) StartServer(opts ServerOptions) error {
if disableSSH, _ := GetEnv("DISABLE_SSH"); disableSSH == "true" {
return errors.New("SSH disabled")
}
if a.server != nil {
return errors.New("server already started")
}
@@ -166,9 +165,20 @@ func (a *Agent) handleSSHRequest(w io.Writer, req *common.HubRequest[cbor.RawMes
}
// responder that writes AgentResponse to stdout
// Uses legacy typed fields for backward compatibility with <= 0.17
sshResponder := func(data any, requestID *uint32) error {
response := newAgentResponse(data, requestID)
response := common.AgentResponse{Id: requestID}
switch v := data.(type) {
case *system.CombinedData:
response.SystemData = v
case string:
response.String = &v
case map[string]smart.SmartData:
response.SmartData = v
case systemd.ServiceDetails:
response.ServiceInfo = v
default:
response.Error = fmt.Sprintf("unsupported response type: %T", data)
}
return cbor.NewEncoder(w).Encode(response)
}
@@ -192,7 +202,7 @@ func (a *Agent) handleSSHRequest(w io.Writer, req *common.HubRequest[cbor.RawMes
// handleLegacyStats serves the legacy one-shot stats payload for older hubs
func (a *Agent) handleLegacyStats(w io.Writer, hubVersion semver.Version) error {
stats := a.gatherStats(common.DataRequestOptions{CacheTimeMs: 60_000})
stats := a.gatherStats(60_000)
return a.writeToSession(w, stats, hubVersion)
}

View File

@@ -1,5 +1,3 @@
//go:build testing
package agent
import (
@@ -182,23 +180,6 @@ func TestStartServer(t *testing.T) {
}
}
func TestStartServerDisableSSH(t *testing.T) {
os.Setenv("BESZEL_AGENT_DISABLE_SSH", "true")
defer os.Unsetenv("BESZEL_AGENT_DISABLE_SSH")
agent, err := NewAgent("")
require.NoError(t, err)
opts := ServerOptions{
Network: "tcp",
Addr: ":45990",
}
err = agent.StartServer(opts)
assert.Error(t, err)
assert.Contains(t, err.Error(), "SSH disabled")
}
/////////////////////////////////////////////////////////////////
//////////////////// ParseKeys Tests ////////////////////////////
/////////////////////////////////////////////////////////////////
@@ -532,7 +513,7 @@ func TestWriteToSessionEncoding(t *testing.T) {
err = json.Unmarshal([]byte(encodedData), &decodedJson)
assert.Error(t, err, "Should not be valid JSON data")
assert.Equal(t, testData.Details.Hostname, decodedCbor.Details.Hostname)
assert.Equal(t, testData.Info.Hostname, decodedCbor.Info.Hostname)
assert.Equal(t, testData.Stats.Cpu, decodedCbor.Stats.Cpu)
} else {
// Should be JSON - try to decode as JSON
@@ -545,7 +526,7 @@ func TestWriteToSessionEncoding(t *testing.T) {
assert.Error(t, err, "Should not be valid CBOR data")
// Verify the decoded JSON data matches our test data
assert.Equal(t, testData.Details.Hostname, decodedJson.Details.Hostname)
assert.Equal(t, testData.Info.Hostname, decodedJson.Info.Hostname)
assert.Equal(t, testData.Stats.Cpu, decodedJson.Stats.Cpu)
// Verify it looks like JSON (starts with '{' and contains readable field names)
@@ -569,12 +550,13 @@ func createTestCombinedData() *system.CombinedData {
DiskUsed: 549755813888, // 512GB
DiskPct: 50.0,
},
Details: &system.Details{
Hostname: "test-host",
},
Info: system.Info{
Hostname: "test-host",
Cores: 8,
CpuModel: "Test CPU Model",
Uptime: 3600,
AgentVersion: "0.12.0",
Os: system.Linux,
},
Containers: []*container.Stats{
{

View File

@@ -8,10 +8,8 @@ import (
"encoding/json"
"errors"
"fmt"
"log/slog"
"os"
"os/exec"
"path/filepath"
"runtime"
"strconv"
"strings"
@@ -19,6 +17,8 @@ import (
"time"
"github.com/henrygd/beszel/internal/entities/smart"
"golang.org/x/exp/slog"
)
// SmartManager manages data collection for SMART devices
@@ -28,7 +28,7 @@ type SmartManager struct {
SmartDevices []*DeviceInfo
refreshMutex sync.Mutex
lastScanTime time.Time
smartctlPath string
binPath string
excludedDevices map[string]struct{}
}
@@ -53,12 +53,6 @@ type DeviceInfo struct {
parserType string
}
// deviceKey is a composite key for a device, used to identify a device uniquely.
type deviceKey struct {
name string
deviceType string
}
var errNoValidSmartData = fmt.Errorf("no valid SMART data found") // Error for missing data
// Refresh updates SMART data for all known devices
@@ -170,35 +164,27 @@ func (sm *SmartManager) ScanDevices(force bool) error {
configuredDevices = parsedDevices
}
ctx, cancel := context.WithTimeout(context.Background(), 5*time.Second)
defer cancel()
cmd := exec.CommandContext(ctx, sm.binPath, "--scan", "-j")
output, err := cmd.Output()
var (
scanErr error
scannedDevices []*DeviceInfo
hasValidScan bool
)
if sm.smartctlPath != "" {
ctx, cancel := context.WithTimeout(context.Background(), 10*time.Second)
defer cancel()
cmd := exec.CommandContext(ctx, sm.smartctlPath, "--scan", "-j")
output, err := cmd.Output()
if err != nil {
scanErr = err
} else {
scannedDevices, hasValidScan = sm.parseScan(output)
if !hasValidScan {
scanErr = errNoValidSmartData
}
if err != nil {
scanErr = err
} else {
scannedDevices, hasValidScan = sm.parseScan(output)
if !hasValidScan {
scanErr = errNoValidSmartData
}
}
// Add eMMC devices (Linux only) by reading sysfs health fields. This does not
// require smartctl and does not scan the whole device.
if emmcDevices := scanEmmcDevices(); len(emmcDevices) > 0 {
scannedDevices = append(scannedDevices, emmcDevices...)
hasValidScan = true
}
finalDevices := mergeDeviceLists(currentDevices, scannedDevices, configuredDevices)
finalDevices = sm.filterExcludedDevices(finalDevices)
sm.updateSmartDevices(finalDevices)
@@ -215,11 +201,7 @@ func (sm *SmartManager) ScanDevices(force bool) error {
}
func (sm *SmartManager) parseConfiguredDevices(config string) ([]*DeviceInfo, error) {
splitChar := os.Getenv("SMART_DEVICES_SEPARATOR")
if splitChar == "" {
splitChar = ","
}
entries := strings.Split(config, splitChar)
entries := strings.Split(config, ",")
devices := make([]*DeviceInfo, 0, len(entries))
for _, entry := range entries {
entry = strings.TrimSpace(entry)
@@ -343,13 +325,6 @@ func normalizeParserType(value string) string {
}
}
// makeDeviceKey creates a composite key from device name and type.
// This allows multiple drives under the same device path (e.g., RAID controllers)
// to be tracked separately.
func makeDeviceKey(name, deviceType string) deviceKey {
return deviceKey{name: name, deviceType: deviceType}
}
// parseSmartOutput attempts each SMART parser, optionally detecting the type when
// it is not provided, and updates the device info when a parser succeeds.
func (sm *SmartManager) parseSmartOutput(deviceInfo *DeviceInfo, output []byte) bool {
@@ -450,75 +425,35 @@ func (sm *SmartManager) CollectSmart(deviceInfo *DeviceInfo) error {
return errNoValidSmartData
}
// eMMC health is not exposed via SMART on Linux, but the kernel provides
// wear / EOL indicators via sysfs. Prefer that path when available.
if deviceInfo != nil {
if ok, err := sm.collectEmmcHealth(deviceInfo); ok {
return err
}
}
if sm.smartctlPath == "" {
return errNoValidSmartData
}
// slog.Info("collecting SMART data", "device", deviceInfo.Name, "type", deviceInfo.Type, "has_existing_data", sm.hasDataForDevice(deviceInfo.Name))
// Check if we have any existing data for this device
hasExistingData := sm.hasDataForDevice(deviceInfo.Name)
ctx, cancel := context.WithTimeout(context.Background(), 15*time.Second)
ctx, cancel := context.WithTimeout(context.Background(), 2*time.Second)
defer cancel()
// Try with -n standby first if we have existing data
args := sm.smartctlArgs(deviceInfo, hasExistingData)
cmd := exec.CommandContext(ctx, sm.smartctlPath, args...)
args := sm.smartctlArgs(deviceInfo, true)
cmd := exec.CommandContext(ctx, sm.binPath, args...)
output, err := cmd.CombinedOutput()
// Check if device is in standby (exit status 2)
if exitErr, ok := errors.AsType[*exec.ExitError](err); ok && exitErr.ExitCode() == 2 {
if exitErr, ok := err.(*exec.ExitError); ok && exitErr.ExitCode() == 2 {
if hasExistingData {
// Device is in standby and we have cached data, keep using cache
return nil
}
// No cached data, need to collect initial data by bypassing standby
ctx2, cancel2 := context.WithTimeout(context.Background(), 15*time.Second)
ctx2, cancel2 := context.WithTimeout(context.Background(), 2*time.Second)
defer cancel2()
args = sm.smartctlArgs(deviceInfo, false)
cmd = exec.CommandContext(ctx2, sm.smartctlPath, args...)
cmd = exec.CommandContext(ctx2, sm.binPath, args...)
output, err = cmd.CombinedOutput()
}
hasValidData := sm.parseSmartOutput(deviceInfo, output)
// If NVMe controller path failed, try namespace path as fallback.
// NVMe controllers (/dev/nvme0) don't always support SMART queries. See github.com/henrygd/beszel/issues/1504
if !hasValidData && err != nil && isNvmeControllerPath(deviceInfo.Name) {
controllerPath := deviceInfo.Name
namespacePath := controllerPath + "n1"
if !sm.isExcludedDevice(namespacePath) {
deviceInfo.Name = namespacePath
ctx3, cancel3 := context.WithTimeout(context.Background(), 15*time.Second)
defer cancel3()
args = sm.smartctlArgs(deviceInfo, false)
cmd = exec.CommandContext(ctx3, sm.smartctlPath, args...)
output, err = cmd.CombinedOutput()
hasValidData = sm.parseSmartOutput(deviceInfo, output)
// Auto-exclude the controller path so future scans don't re-add it
if hasValidData {
sm.Lock()
if sm.excludedDevices == nil {
sm.excludedDevices = make(map[string]struct{})
}
sm.excludedDevices[controllerPath] = struct{}{}
sm.Unlock()
slog.Debug("auto-excluded NVMe controller path", "path", controllerPath)
}
}
}
if !hasValidData {
if err != nil {
slog.Info("smartctl failed", "device", deviceInfo.Name, "err", err)
@@ -534,12 +469,10 @@ func (sm *SmartManager) CollectSmart(deviceInfo *DeviceInfo) error {
// smartctlArgs returns the arguments for the smartctl command
// based on the device type and whether to include standby mode
func (sm *SmartManager) smartctlArgs(deviceInfo *DeviceInfo, includeStandby bool) []string {
args := make([]string, 0, 9)
var deviceType, parserType string
args := make([]string, 0, 7)
if deviceInfo != nil {
deviceType = strings.ToLower(deviceInfo.Type)
parserType = strings.ToLower(deviceInfo.parserType)
deviceType := strings.ToLower(deviceInfo.Type)
// types sometimes misidentified in scan; see github.com/henrygd/beszel/issues/1345
if deviceType != "" && deviceType != "scsi" && deviceType != "ata" {
args = append(args, "-d", deviceInfo.Type)
@@ -547,13 +480,6 @@ func (sm *SmartManager) smartctlArgs(deviceInfo *DeviceInfo, includeStandby bool
}
args = append(args, "-a", "--json=c")
effectiveType := parserType
if effectiveType == "" {
effectiveType = deviceType
}
if effectiveType == "sat" || effectiveType == "ata" {
args = append(args, "-l", "devstat")
}
if includeStandby {
args = append(args, "-n", "standby")
@@ -614,28 +540,6 @@ func mergeDeviceLists(existing, scanned, configured []*DeviceInfo) []*DeviceInfo
return existing
}
// buildUniqueNameIndex returns devices that appear exactly once by name.
// It is used to safely apply name-only fallbacks without RAID ambiguity.
buildUniqueNameIndex := func(devices []*DeviceInfo) map[string]*DeviceInfo {
counts := make(map[string]int, len(devices))
for _, dev := range devices {
if dev == nil || dev.Name == "" {
continue
}
counts[dev.Name]++
}
unique := make(map[string]*DeviceInfo, len(counts))
for _, dev := range devices {
if dev == nil || dev.Name == "" {
continue
}
if counts[dev.Name] == 1 {
unique[dev.Name] = dev
}
}
return unique
}
// preserveVerifiedType copies the verified type/parser metadata from an existing
// device record so that subsequent scans/config updates never downgrade a
// previously verified device.
@@ -648,90 +552,69 @@ func mergeDeviceLists(existing, scanned, configured []*DeviceInfo) []*DeviceInfo
target.parserType = prev.parserType
}
// applyConfiguredMetadata updates a matched device with any configured
// overrides, preserving verified type data when present.
applyConfiguredMetadata := func(existingDev, configuredDev *DeviceInfo) {
// Only update the type if it has not been verified yet; otherwise we
// keep the existing verified metadata intact.
if configuredDev.Type != "" && !existingDev.typeVerified {
newType := strings.TrimSpace(configuredDev.Type)
existingDev.Type = newType
existingDev.typeVerified = false
existingDev.parserType = normalizeParserType(newType)
}
if configuredDev.InfoName != "" {
existingDev.InfoName = configuredDev.InfoName
}
if configuredDev.Protocol != "" {
existingDev.Protocol = configuredDev.Protocol
}
}
existingIndex := make(map[deviceKey]*DeviceInfo, len(existing))
existingIndex := make(map[string]*DeviceInfo, len(existing))
for _, dev := range existing {
if dev == nil || dev.Name == "" {
continue
}
existingIndex[makeDeviceKey(dev.Name, dev.Type)] = dev
existingIndex[dev.Name] = dev
}
existingByName := buildUniqueNameIndex(existing)
finalDevices := make([]*DeviceInfo, 0, len(scanned)+len(configured))
deviceIndex := make(map[deviceKey]*DeviceInfo, len(scanned)+len(configured))
deviceIndex := make(map[string]*DeviceInfo, len(scanned)+len(configured))
// Start with the newly scanned devices so we always surface fresh metadata,
// but ensure we retain any previously verified parser assignment.
for _, scannedDevice := range scanned {
if scannedDevice == nil || scannedDevice.Name == "" {
for _, dev := range scanned {
if dev == nil || dev.Name == "" {
continue
}
// Work on a copy so we can safely adjust metadata without mutating the
// input slices that may be reused elsewhere.
copyDev := *scannedDevice
key := makeDeviceKey(copyDev.Name, copyDev.Type)
if prev := existingIndex[key]; prev != nil {
preserveVerifiedType(&copyDev, prev)
} else if prev := existingByName[copyDev.Name]; prev != nil {
copyDev := *dev
if prev := existingIndex[copyDev.Name]; prev != nil {
preserveVerifiedType(&copyDev, prev)
}
finalDevices = append(finalDevices, &copyDev)
copyKey := makeDeviceKey(copyDev.Name, copyDev.Type)
deviceIndex[copyKey] = finalDevices[len(finalDevices)-1]
deviceIndex[copyDev.Name] = finalDevices[len(finalDevices)-1]
}
deviceIndexByName := buildUniqueNameIndex(finalDevices)
// Merge configured devices on top so users can override scan results (except
// for verified type information).
for _, configuredDevice := range configured {
if configuredDevice == nil || configuredDevice.Name == "" {
for _, dev := range configured {
if dev == nil || dev.Name == "" {
continue
}
key := makeDeviceKey(configuredDevice.Name, configuredDevice.Type)
if existingDev, ok := deviceIndex[key]; ok {
applyConfiguredMetadata(existingDev, configuredDevice)
continue
}
if existingDev := deviceIndexByName[configuredDevice.Name]; existingDev != nil {
applyConfiguredMetadata(existingDev, configuredDevice)
if existingDev, ok := deviceIndex[dev.Name]; ok {
// Only update the type if it has not been verified yet; otherwise we
// keep the existing verified metadata intact.
if dev.Type != "" && !existingDev.typeVerified {
newType := strings.TrimSpace(dev.Type)
existingDev.Type = newType
existingDev.typeVerified = false
existingDev.parserType = normalizeParserType(newType)
}
if dev.InfoName != "" {
existingDev.InfoName = dev.InfoName
}
if dev.Protocol != "" {
existingDev.Protocol = dev.Protocol
}
continue
}
copyDev := *configuredDevice
key = makeDeviceKey(copyDev.Name, copyDev.Type)
if prev := existingIndex[key]; prev != nil {
preserveVerifiedType(&copyDev, prev)
} else if prev := existingByName[copyDev.Name]; prev != nil {
copyDev := *dev
if prev := existingIndex[copyDev.Name]; prev != nil {
preserveVerifiedType(&copyDev, prev)
} else if copyDev.Type != "" {
copyDev.parserType = normalizeParserType(copyDev.Type)
}
finalDevices = append(finalDevices, &copyDev)
copyKey := makeDeviceKey(copyDev.Name, copyDev.Type)
deviceIndex[copyKey] = finalDevices[len(finalDevices)-1]
deviceIndex[copyDev.Name] = finalDevices[len(finalDevices)-1]
}
return finalDevices
@@ -749,14 +632,12 @@ func (sm *SmartManager) updateSmartDevices(devices []*DeviceInfo) {
return
}
validKeys := make(map[deviceKey]struct{}, len(devices))
nameCounts := make(map[string]int, len(devices))
validNames := make(map[string]struct{}, len(devices))
for _, device := range devices {
if device == nil || device.Name == "" {
continue
}
validKeys[makeDeviceKey(device.Name, device.Type)] = struct{}{}
nameCounts[device.Name]++
validNames[device.Name] = struct{}{}
}
for key, data := range sm.SmartDataMap {
@@ -765,11 +646,7 @@ func (sm *SmartManager) updateSmartDevices(devices []*DeviceInfo) {
continue
}
if data.DiskType == "" {
if nameCounts[data.DiskName] == 1 {
continue
}
} else if _, ok := validKeys[makeDeviceKey(data.DiskName, data.DiskType)]; ok {
if _, ok := validNames[data.DiskName]; ok {
continue
}
@@ -857,11 +734,6 @@ func (sm *SmartManager) parseSmartForSata(output []byte) (bool, int) {
smartData.FirmwareVersion = data.FirmwareVersion
smartData.Capacity = data.UserCapacity.Bytes
smartData.Temperature = data.Temperature.Current
if smartData.Temperature == 0 {
if temp, ok := temperatureFromAtaDeviceStatistics(data.AtaDeviceStatistics); ok {
smartData.Temperature = temp
}
}
smartData.SmartStatus = getSmartStatus(smartData.Temperature, data.SmartStatus.Passed)
smartData.DiskName = data.Device.Name
smartData.DiskType = data.Device.Type
@@ -900,36 +772,6 @@ func getSmartStatus(temperature uint8, passed bool) string {
}
}
func temperatureFromAtaDeviceStatistics(stats smart.AtaDeviceStatistics) (uint8, bool) {
entry := findAtaDeviceStatisticsEntry(stats, 5, "Current Temperature")
if entry == nil || entry.Value == nil {
return 0, false
}
if *entry.Value > 255 {
return 0, false
}
return uint8(*entry.Value), true
}
// findAtaDeviceStatisticsEntry centralizes ATA devstat lookups so additional
// metrics can be pulled from the same structure in the future.
func findAtaDeviceStatisticsEntry(stats smart.AtaDeviceStatistics, pageNumber uint8, entryName string) *smart.AtaDeviceStatisticsEntry {
for pageIdx := range stats.Pages {
page := &stats.Pages[pageIdx]
if page.Number != pageNumber {
continue
}
for entryIdx := range page.Table {
entry := &page.Table[entryIdx]
if !strings.EqualFold(entry.Name, entryName) {
continue
}
return entry
}
}
return nil
}
func (sm *SmartManager) parseSmartForScsi(output []byte) (bool, int) {
var data smart.SmartInfoForScsi
@@ -1115,27 +957,6 @@ func (sm *SmartManager) detectSmartctl() (string, error) {
return "", errors.New("smartctl not found")
}
// isNvmeControllerPath checks if the path matches an NVMe controller pattern
// like /dev/nvme0, /dev/nvme1, etc. (without namespace suffix like n1)
func isNvmeControllerPath(path string) bool {
base := filepath.Base(path)
if !strings.HasPrefix(base, "nvme") {
return false
}
suffix := strings.TrimPrefix(base, "nvme")
if suffix == "" {
return false
}
// Controller paths are just "nvme" + digits (e.g., nvme0, nvme1)
// Namespace paths have "n" after the controller number (e.g., nvme0n1)
for _, c := range suffix {
if c < '0' || c > '9' {
return false
}
}
return true
}
// NewSmartManager creates and initializes a new SmartManager
func NewSmartManager() (*SmartManager, error) {
sm := &SmartManager{
@@ -1143,15 +964,11 @@ func NewSmartManager() (*SmartManager, error) {
}
sm.refreshExcludedDevices()
path, err := sm.detectSmartctl()
slog.Debug("smartctl", "path", path, "err", err)
if err != nil {
// Keep the previous fail-fast behavior unless this Linux host exposes
// eMMC health via sysfs, in which case smartctl is optional.
if runtime.GOOS == "linux" && len(scanEmmcDevices()) > 0 {
return sm, nil
}
slog.Debug(err.Error())
return nil, err
}
sm.smartctlPath = path
slog.Debug("smartctl", "path", path)
sm.binPath = path
return sm, nil
}

View File

@@ -1,4 +1,5 @@
//go:build testing
// +build testing
package agent
@@ -88,39 +89,6 @@ func TestParseSmartForSata(t *testing.T) {
}
}
func TestParseSmartForSataDeviceStatisticsTemperature(t *testing.T) {
jsonPayload := []byte(`{
"smartctl": {"exit_status": 0},
"device": {"name": "/dev/sdb", "type": "sat"},
"model_name": "SanDisk SSD U110 16GB",
"serial_number": "DEVSTAT123",
"firmware_version": "U21B001",
"user_capacity": {"bytes": 16013942784},
"smart_status": {"passed": true},
"ata_smart_attributes": {"table": []},
"ata_device_statistics": {
"pages": [
{
"number": 5,
"name": "Temperature Statistics",
"table": [
{"name": "Current Temperature", "value": 22, "flags": {"valid": true}}
]
}
]
}
}`)
sm := &SmartManager{SmartDataMap: make(map[string]*smart.SmartData)}
hasData, exitStatus := sm.parseSmartForSata(jsonPayload)
require.True(t, hasData)
assert.Equal(t, 0, exitStatus)
deviceData, ok := sm.SmartDataMap["DEVSTAT123"]
require.True(t, ok, "expected smart data entry for serial DEVSTAT123")
assert.Equal(t, uint8(22), deviceData.Temperature)
}
func TestParseSmartForSataParentheticalRawValue(t *testing.T) {
jsonPayload := []byte(`{
"smartctl": {"exit_status": 0},
@@ -227,24 +195,6 @@ func TestDevicesSnapshotReturnsCopy(t *testing.T) {
assert.Len(t, snapshot, 2)
}
func TestScanDevicesWithEnvOverrideAndSeparator(t *testing.T) {
t.Setenv("SMART_DEVICES_SEPARATOR", "|")
t.Setenv("SMART_DEVICES", "/dev/sda:jmb39x-q,0|/dev/nvme0:nvme")
sm := &SmartManager{
SmartDataMap: make(map[string]*smart.SmartData),
}
err := sm.ScanDevices(true)
require.NoError(t, err)
require.Len(t, sm.SmartDevices, 2)
assert.Equal(t, "/dev/sda", sm.SmartDevices[0].Name)
assert.Equal(t, "jmb39x-q,0", sm.SmartDevices[0].Type)
assert.Equal(t, "/dev/nvme0", sm.SmartDevices[1].Name)
assert.Equal(t, "nvme", sm.SmartDevices[1].Type)
}
func TestScanDevicesWithEnvOverride(t *testing.T) {
t.Setenv("SMART_DEVICES", "/dev/sda:sat, /dev/nvme0:nvme")
@@ -299,21 +249,15 @@ func TestSmartctlArgs(t *testing.T) {
sataDevice := &DeviceInfo{Name: "/dev/sda", Type: "sat"}
assert.Equal(t,
[]string{"-d", "sat", "-a", "--json=c", "-l", "devstat", "-n", "standby", "/dev/sda"},
[]string{"-d", "sat", "-a", "--json=c", "-n", "standby", "/dev/sda"},
sm.smartctlArgs(sataDevice, true),
)
assert.Equal(t,
[]string{"-d", "sat", "-a", "--json=c", "-l", "devstat", "/dev/sda"},
[]string{"-d", "sat", "-a", "--json=c", "/dev/sda"},
sm.smartctlArgs(sataDevice, false),
)
nvmeDevice := &DeviceInfo{Name: "/dev/nvme0", Type: "nvme"}
assert.Equal(t,
[]string{"-d", "nvme", "-a", "--json=c", "-n", "standby", "/dev/nvme0"},
sm.smartctlArgs(nvmeDevice, true),
)
assert.Equal(t,
[]string{"-a", "--json=c", "-n", "standby"},
sm.smartctlArgs(nil, true),
@@ -498,88 +442,6 @@ func TestMergeDeviceListsUpdatesTypeWhenUnverified(t *testing.T) {
assert.Equal(t, "", device.parserType)
}
func TestMergeDeviceListsHandlesDevicesWithSameNameAndDifferentTypes(t *testing.T) {
// There are use cases where the same device name is re-used,
// for example, a RAID controller with multiple drives.
scanned := []*DeviceInfo{
{Name: "/dev/sda", Type: "megaraid,0"},
{Name: "/dev/sda", Type: "megaraid,1"},
{Name: "/dev/sda", Type: "megaraid,2"},
}
merged := mergeDeviceLists(nil, scanned, nil)
require.Len(t, merged, 3, "should have 3 separate devices for RAID controller")
byKey := make(map[string]*DeviceInfo, len(merged))
for _, dev := range merged {
key := dev.Name + "|" + dev.Type
byKey[key] = dev
}
assert.Contains(t, byKey, "/dev/sda|megaraid,0")
assert.Contains(t, byKey, "/dev/sda|megaraid,1")
assert.Contains(t, byKey, "/dev/sda|megaraid,2")
}
func TestMergeDeviceListsHandlesMixedRAIDAndRegular(t *testing.T) {
// Test mixing RAID drives with regular devices
scanned := []*DeviceInfo{
{Name: "/dev/sda", Type: "megaraid,0"},
{Name: "/dev/sda", Type: "megaraid,1"},
{Name: "/dev/sdb", Type: "sat"},
{Name: "/dev/nvme0", Type: "nvme"},
}
merged := mergeDeviceLists(nil, scanned, nil)
require.Len(t, merged, 4, "should have 4 separate devices")
byKey := make(map[string]*DeviceInfo, len(merged))
for _, dev := range merged {
key := dev.Name + "|" + dev.Type
byKey[key] = dev
}
assert.Contains(t, byKey, "/dev/sda|megaraid,0")
assert.Contains(t, byKey, "/dev/sda|megaraid,1")
assert.Contains(t, byKey, "/dev/sdb|sat")
assert.Contains(t, byKey, "/dev/nvme0|nvme")
}
func TestUpdateSmartDevicesPreservesRAIDDrives(t *testing.T) {
// Test that updateSmartDevices correctly validates RAID drives using composite keys
sm := &SmartManager{
SmartDevices: []*DeviceInfo{
{Name: "/dev/sda", Type: "megaraid,0"},
{Name: "/dev/sda", Type: "megaraid,1"},
},
SmartDataMap: map[string]*smart.SmartData{
"serial-0": {
DiskName: "/dev/sda",
DiskType: "megaraid,0",
SerialNumber: "serial-0",
},
"serial-1": {
DiskName: "/dev/sda",
DiskType: "megaraid,1",
SerialNumber: "serial-1",
},
"serial-stale": {
DiskName: "/dev/sda",
DiskType: "megaraid,2",
SerialNumber: "serial-stale",
},
},
}
sm.updateSmartDevices(sm.SmartDevices)
// serial-0 and serial-1 should be preserved (matching devices exist)
assert.Contains(t, sm.SmartDataMap, "serial-0")
assert.Contains(t, sm.SmartDataMap, "serial-1")
// serial-stale should be removed (no matching device)
assert.NotContains(t, sm.SmartDataMap, "serial-stale")
}
func TestParseSmartOutputMarksVerified(t *testing.T) {
fixturePath := filepath.Join("test-data", "smart", "nvme0.json")
data, err := os.ReadFile(fixturePath)
@@ -918,36 +780,3 @@ func TestFilterExcludedDevices(t *testing.T) {
})
}
}
func TestIsNvmeControllerPath(t *testing.T) {
tests := []struct {
path string
expected bool
}{
// Controller paths (should return true)
{"/dev/nvme0", true},
{"/dev/nvme1", true},
{"/dev/nvme10", true},
{"nvme0", true},
// Namespace paths (should return false)
{"/dev/nvme0n1", false},
{"/dev/nvme1n1", false},
{"/dev/nvme0n1p1", false},
{"nvme0n1", false},
// Non-NVMe paths (should return false)
{"/dev/sda", false},
{"/dev/sda1", false},
{"/dev/hda", false},
{"", false},
{"/dev/nvme", false},
}
for _, tt := range tests {
t.Run(tt.path, func(t *testing.T) {
result := isNvmeControllerPath(tt.path)
assert.Equal(t, tt.expected, result, "path: %s", tt.path)
})
}
}

View File

@@ -2,18 +2,15 @@ package agent
import (
"bufio"
"errors"
"fmt"
"log/slog"
"os"
"runtime"
"strconv"
"strings"
"time"
"github.com/henrygd/beszel"
"github.com/henrygd/beszel/agent/battery"
"github.com/henrygd/beszel/internal/entities/container"
"github.com/henrygd/beszel/internal/entities/system"
"github.com/shirou/gopsutil/v4/cpu"
@@ -30,79 +27,41 @@ type prevDisk struct {
}
// Sets initial / non-changing values about the host system
func (a *Agent) refreshSystemDetails() {
func (a *Agent) initializeSystemInfo() {
a.systemInfo.AgentVersion = beszel.Version
// get host info from Docker if available
var hostInfo container.HostInfo
if a.dockerManager != nil {
a.systemDetails.Podman = a.dockerManager.IsPodman()
hostInfo, _ = a.dockerManager.GetHostInfo()
}
a.systemDetails.Hostname, _ = os.Hostname()
if arch, err := host.KernelArch(); err == nil {
a.systemDetails.Arch = arch
} else {
a.systemDetails.Arch = runtime.GOARCH
}
a.systemInfo.Hostname, _ = os.Hostname()
platform, _, version, _ := host.PlatformInformation()
if platform == "darwin" {
a.systemDetails.Os = system.Darwin
a.systemDetails.OsName = fmt.Sprintf("macOS %s", version)
a.systemInfo.KernelVersion = version
a.systemInfo.Os = system.Darwin
} else if strings.Contains(platform, "indows") {
a.systemDetails.Os = system.Windows
a.systemDetails.OsName = strings.Replace(platform, "Microsoft ", "", 1)
a.systemDetails.Kernel = version
a.systemInfo.KernelVersion = fmt.Sprintf("%s %s", strings.Replace(platform, "Microsoft ", "", 1), version)
a.systemInfo.Os = system.Windows
} else if platform == "freebsd" {
a.systemDetails.Os = system.Freebsd
a.systemDetails.Kernel, _ = host.KernelVersion()
if prettyName, err := getOsPrettyName(); err == nil {
a.systemDetails.OsName = prettyName
} else {
a.systemDetails.OsName = "FreeBSD"
}
a.systemInfo.Os = system.Freebsd
a.systemInfo.KernelVersion = version
} else {
a.systemDetails.Os = system.Linux
a.systemDetails.OsName = hostInfo.OperatingSystem
if a.systemDetails.OsName == "" {
if prettyName, err := getOsPrettyName(); err == nil {
a.systemDetails.OsName = prettyName
} else {
a.systemDetails.OsName = platform
}
}
a.systemDetails.Kernel = hostInfo.KernelVersion
if a.systemDetails.Kernel == "" {
a.systemDetails.Kernel, _ = host.KernelVersion()
}
a.systemInfo.Os = system.Linux
}
if a.systemInfo.KernelVersion == "" {
a.systemInfo.KernelVersion, _ = host.KernelVersion()
}
// cpu model
if info, err := cpu.Info(); err == nil && len(info) > 0 {
a.systemDetails.CpuModel = info[0].ModelName
a.systemInfo.CpuModel = info[0].ModelName
}
// cores / threads
cores, _ := cpu.Counts(false)
threads := hostInfo.NCPU
if threads == 0 {
threads, _ = cpu.Counts(true)
}
// in lxc, logical cores reflects container limits, so use that as cores if lower
if threads > 0 && threads < cores {
cores = threads
}
a.systemDetails.Cores = cores
a.systemDetails.Threads = threads
// total memory
a.systemDetails.MemoryTotal = hostInfo.MemTotal
if a.systemDetails.MemoryTotal == 0 {
if v, err := mem.VirtualMemory(); err == nil {
a.systemDetails.MemoryTotal = v.Total
a.systemInfo.Cores, _ = cpu.Counts(false)
if threads, err := cpu.Counts(true); err == nil {
if threads > 0 && threads < a.systemInfo.Cores {
// in lxc logical cores reflects container limits, so use that as cores if lower
a.systemInfo.Cores = threads
} else {
a.systemInfo.Threads = threads
}
}
@@ -236,16 +195,21 @@ func (a *Agent) getSystemStats(cacheTimeMs uint16) system.Stats {
}
}
// update system info
// update base system info
a.systemInfo.ConnectionType = a.connectionManager.ConnectionType
a.systemInfo.Cpu = systemStats.Cpu
a.systemInfo.LoadAvg = systemStats.LoadAvg
// TODO: remove these in future release in favor of load avg array
a.systemInfo.LoadAvg1 = systemStats.LoadAvg[0]
a.systemInfo.LoadAvg5 = systemStats.LoadAvg[1]
a.systemInfo.LoadAvg15 = systemStats.LoadAvg[2]
a.systemInfo.MemPct = systemStats.MemPct
a.systemInfo.DiskPct = systemStats.DiskPct
a.systemInfo.Battery = systemStats.Battery
a.systemInfo.Uptime, _ = host.Uptime()
// TODO: in future release, remove MB bandwidth values in favor of bytes
a.systemInfo.Bandwidth = twoDecimals(systemStats.NetworkSent + systemStats.NetworkRecv)
a.systemInfo.BandwidthBytes = systemStats.Bandwidth[0] + systemStats.Bandwidth[1]
a.systemInfo.Threads = a.systemDetails.Threads
slog.Debug("sysinfo", "data", a.systemInfo)
return systemStats
}
@@ -275,24 +239,3 @@ func getARCSize() (uint64, error) {
return 0, fmt.Errorf("failed to parse size field")
}
// getOsPrettyName attempts to get the pretty OS name from /etc/os-release on Linux systems
func getOsPrettyName() (string, error) {
file, err := os.Open("/etc/os-release")
if err != nil {
return "", err
}
defer file.Close()
scanner := bufio.NewScanner(file)
for scanner.Scan() {
line := scanner.Text()
if after, ok := strings.CutPrefix(line, "PRETTY_NAME="); ok {
value := after
value = strings.Trim(value, `"`)
return value, nil
}
}
return "", errors.New("pretty name not found")
}

View File

@@ -8,7 +8,6 @@ import (
"log/slog"
"maps"
"math"
"os"
"strconv"
"strings"
"sync"
@@ -29,36 +28,11 @@ type systemdManager struct {
patterns []string
}
// isSystemdAvailable checks if systemd is used on the system to avoid unnecessary connection attempts (#1548)
func isSystemdAvailable() bool {
paths := []string{
"/run/systemd/system",
"/run/dbus/system_bus_socket",
"/var/run/dbus/system_bus_socket",
}
for _, path := range paths {
if _, err := os.Stat(path); err == nil {
return true
}
}
if data, err := os.ReadFile("/proc/1/comm"); err == nil {
return strings.TrimSpace(string(data)) == "systemd"
}
return false
}
// newSystemdManager creates a new systemdManager.
func newSystemdManager() (*systemdManager, error) {
if skipSystemd, _ := GetEnv("SKIP_SYSTEMD"); skipSystemd == "true" {
return nil, nil
}
// Check if systemd is available on the system before attempting connection
if !isSystemdAvailable() {
slog.Debug("Systemd not available")
return nil, nil
}
conn, err := dbus.NewSystemConnectionContext(context.Background())
if err != nil {
slog.Debug("Error connecting to systemd", "err", err, "ref", "https://beszel.dev/guide/systemd")
@@ -144,27 +118,13 @@ func (sm *systemdManager) getServiceStats(conn *dbus.Conn, refresh bool) []*syst
return nil
}
// Track which units are currently present to remove stale entries
currentUnits := make(map[string]struct{}, len(units))
for _, unit := range units {
currentUnits[unit.Name] = struct{}{}
service, err := sm.updateServiceStats(conn, unit)
if err != nil {
continue
}
services = append(services, service)
}
// Remove services that no longer exist in systemd
sm.Lock()
for unitName := range sm.serviceStatsMap {
if _, exists := currentUnits[unitName]; !exists {
delete(sm.serviceStatsMap, unitName)
}
}
sm.Unlock()
sm.hasFreshStats = true
return services
}

View File

@@ -19,11 +19,11 @@ func TestSystemdManagerGetServiceStats(t *testing.T) {
assert.NoError(t, err)
// Test with refresh = true
result := manager.getServiceStats("any-service", true)
result := manager.getServiceStats(true)
assert.Nil(t, result)
// Test with refresh = false
result = manager.getServiceStats("any-service", false)
result = manager.getServiceStats(false)
assert.Nil(t, result)
}

View File

@@ -4,7 +4,6 @@ package agent
import (
"os"
"strings"
"testing"
"github.com/stretchr/testify/assert"
@@ -49,35 +48,6 @@ func TestUnescapeServiceNameInvalid(t *testing.T) {
}
}
func TestIsSystemdAvailable(t *testing.T) {
// Note: This test's result will vary based on the actual system running the tests
// On systems with systemd, it should return true
// On systems without systemd, it should return false
result := isSystemdAvailable()
// Check if either the /run/systemd/system directory exists or PID 1 is systemd
runSystemdExists := false
if _, err := os.Stat("/run/systemd/system"); err == nil {
runSystemdExists = true
}
pid1IsSystemd := false
if data, err := os.ReadFile("/proc/1/comm"); err == nil {
pid1IsSystemd = strings.TrimSpace(string(data)) == "systemd"
}
expected := runSystemdExists || pid1IsSystemd
assert.Equal(t, expected, result, "isSystemdAvailable should correctly detect systemd presence")
// Log the result for informational purposes
if result {
t.Log("Systemd is available on this system")
} else {
t.Log("Systemd is not available on this system")
}
}
func TestGetServicePatterns(t *testing.T) {
tests := []struct {
name string

View File

@@ -1,700 +0,0 @@
# List of AMDGPU IDs
#
# Syntax:
# device_id, revision_id, product_name <-- single tab after comma
1.0.0
1114, C2, AMD Radeon 860M Graphics
1114, C3, AMD Radeon 840M Graphics
1114, D2, AMD Radeon 860M Graphics
1114, D3, AMD Radeon 840M Graphics
1309, 00, AMD Radeon R7 Graphics
130A, 00, AMD Radeon R6 Graphics
130B, 00, AMD Radeon R4 Graphics
130C, 00, AMD Radeon R7 Graphics
130D, 00, AMD Radeon R6 Graphics
130E, 00, AMD Radeon R5 Graphics
130F, 00, AMD Radeon R7 Graphics
130F, D4, AMD Radeon R7 Graphics
130F, D5, AMD Radeon R7 Graphics
130F, D6, AMD Radeon R7 Graphics
130F, D7, AMD Radeon R7 Graphics
1313, 00, AMD Radeon R7 Graphics
1313, D4, AMD Radeon R7 Graphics
1313, D5, AMD Radeon R7 Graphics
1313, D6, AMD Radeon R7 Graphics
1315, 00, AMD Radeon R5 Graphics
1315, D4, AMD Radeon R5 Graphics
1315, D5, AMD Radeon R5 Graphics
1315, D6, AMD Radeon R5 Graphics
1315, D7, AMD Radeon R5 Graphics
1316, 00, AMD Radeon R5 Graphics
1318, 00, AMD Radeon R5 Graphics
131B, 00, AMD Radeon R4 Graphics
131C, 00, AMD Radeon R7 Graphics
131D, 00, AMD Radeon R6 Graphics
1435, AE, AMD Custom GPU 0932
1506, C1, AMD Radeon 610M
1506, C2, AMD Radeon 610M
1506, C3, AMD Radeon 610M
1506, C4, AMD Radeon 610M
150E, C1, AMD Radeon 890M Graphics
150E, C4, AMD Radeon 890M Graphics
150E, C5, AMD Radeon 890M Graphics
150E, C6, AMD Radeon 890M Graphics
150E, D1, AMD Radeon 890M Graphics
150E, D2, AMD Radeon 890M Graphics
150E, D3, AMD Radeon 890M Graphics
1586, C1, Radeon 8060S Graphics
1586, C2, Radeon 8050S Graphics
1586, C4, Radeon 8050S Graphics
1586, D1, Radeon 8060S Graphics
1586, D2, Radeon 8050S Graphics
1586, D4, Radeon 8050S Graphics
1586, D5, Radeon 8040S Graphics
15BF, 00, AMD Radeon 780M Graphics
15BF, 01, AMD Radeon 760M Graphics
15BF, 02, AMD Radeon 780M Graphics
15BF, 03, AMD Radeon 760M Graphics
15BF, C1, AMD Radeon 780M Graphics
15BF, C2, AMD Radeon 780M Graphics
15BF, C3, AMD Radeon 760M Graphics
15BF, C4, AMD Radeon 780M Graphics
15BF, C5, AMD Radeon 740M Graphics
15BF, C6, AMD Radeon 780M Graphics
15BF, C7, AMD Radeon 780M Graphics
15BF, C8, AMD Radeon 760M Graphics
15BF, C9, AMD Radeon 780M Graphics
15BF, CA, AMD Radeon 740M Graphics
15BF, CB, AMD Radeon 760M Graphics
15BF, CC, AMD Radeon 740M Graphics
15BF, CD, AMD Radeon 760M Graphics
15BF, CF, AMD Radeon 780M Graphics
15BF, D0, AMD Radeon 780M Graphics
15BF, D1, AMD Radeon 780M Graphics
15BF, D2, AMD Radeon 780M Graphics
15BF, D3, AMD Radeon 780M Graphics
15BF, D4, AMD Radeon 780M Graphics
15BF, D5, AMD Radeon 760M Graphics
15BF, D6, AMD Radeon 760M Graphics
15BF, D7, AMD Radeon 780M Graphics
15BF, D8, AMD Radeon 740M Graphics
15BF, D9, AMD Radeon 780M Graphics
15BF, DA, AMD Radeon 780M Graphics
15BF, DB, AMD Radeon 760M Graphics
15BF, DC, AMD Radeon 760M Graphics
15BF, DD, AMD Radeon 780M Graphics
15BF, DE, AMD Radeon 740M Graphics
15BF, DF, AMD Radeon 760M Graphics
15BF, F0, AMD Radeon 760M Graphics
15C8, C1, AMD Radeon 740M Graphics
15C8, C2, AMD Radeon 740M Graphics
15C8, C3, AMD Radeon 740M Graphics
15C8, C4, AMD Radeon 740M Graphics
15C8, D1, AMD Radeon 740M Graphics
15C8, D2, AMD Radeon 740M Graphics
15C8, D3, AMD Radeon 740M Graphics
15C8, D4, AMD Radeon 740M Graphics
15D8, 00, AMD Radeon RX Vega 8 Graphics WS
15D8, 91, AMD Radeon Vega 3 Graphics
15D8, 91, AMD Ryzen Embedded R1606G with Radeon Vega Gfx
15D8, 92, AMD Radeon Vega 3 Graphics
15D8, 92, AMD Ryzen Embedded R1505G with Radeon Vega Gfx
15D8, 93, AMD Radeon Vega 1 Graphics
15D8, A1, AMD Radeon Vega 10 Graphics
15D8, A2, AMD Radeon Vega 8 Graphics
15D8, A3, AMD Radeon Vega 6 Graphics
15D8, A4, AMD Radeon Vega 3 Graphics
15D8, B1, AMD Radeon Vega 10 Graphics
15D8, B2, AMD Radeon Vega 8 Graphics
15D8, B3, AMD Radeon Vega 6 Graphics
15D8, B4, AMD Radeon Vega 3 Graphics
15D8, C1, AMD Radeon Vega 10 Graphics
15D8, C2, AMD Radeon Vega 8 Graphics
15D8, C3, AMD Radeon Vega 6 Graphics
15D8, C4, AMD Radeon Vega 3 Graphics
15D8, C5, AMD Radeon Vega 3 Graphics
15D8, C8, AMD Radeon Vega 11 Graphics
15D8, C9, AMD Radeon Vega 8 Graphics
15D8, CA, AMD Radeon Vega 11 Graphics
15D8, CB, AMD Radeon Vega 8 Graphics
15D8, CC, AMD Radeon Vega 3 Graphics
15D8, CE, AMD Radeon Vega 3 Graphics
15D8, CF, AMD Ryzen Embedded R1305G with Radeon Vega Gfx
15D8, D1, AMD Radeon Vega 10 Graphics
15D8, D2, AMD Radeon Vega 8 Graphics
15D8, D3, AMD Radeon Vega 6 Graphics
15D8, D4, AMD Radeon Vega 3 Graphics
15D8, D8, AMD Radeon Vega 11 Graphics
15D8, D9, AMD Radeon Vega 8 Graphics
15D8, DA, AMD Radeon Vega 11 Graphics
15D8, DB, AMD Radeon Vega 3 Graphics
15D8, DB, AMD Radeon Vega 8 Graphics
15D8, DC, AMD Radeon Vega 3 Graphics
15D8, DD, AMD Radeon Vega 3 Graphics
15D8, DE, AMD Radeon Vega 3 Graphics
15D8, DF, AMD Radeon Vega 3 Graphics
15D8, E3, AMD Radeon Vega 3 Graphics
15D8, E4, AMD Ryzen Embedded R1102G with Radeon Vega Gfx
15DD, 81, AMD Ryzen Embedded V1807B with Radeon Vega Gfx
15DD, 82, AMD Ryzen Embedded V1756B with Radeon Vega Gfx
15DD, 83, AMD Ryzen Embedded V1605B with Radeon Vega Gfx
15DD, 84, AMD Radeon Vega 6 Graphics
15DD, 85, AMD Ryzen Embedded V1202B with Radeon Vega Gfx
15DD, 86, AMD Radeon Vega 11 Graphics
15DD, 88, AMD Radeon Vega 8 Graphics
15DD, C1, AMD Radeon Vega 11 Graphics
15DD, C2, AMD Radeon Vega 8 Graphics
15DD, C3, AMD Radeon Vega 3 / 10 Graphics
15DD, C4, AMD Radeon Vega 8 Graphics
15DD, C5, AMD Radeon Vega 3 Graphics
15DD, C6, AMD Radeon Vega 11 Graphics
15DD, C8, AMD Radeon Vega 8 Graphics
15DD, C9, AMD Radeon Vega 11 Graphics
15DD, CA, AMD Radeon Vega 8 Graphics
15DD, CB, AMD Radeon Vega 3 Graphics
15DD, CC, AMD Radeon Vega 6 Graphics
15DD, CE, AMD Radeon Vega 3 Graphics
15DD, CF, AMD Radeon Vega 3 Graphics
15DD, D0, AMD Radeon Vega 10 Graphics
15DD, D1, AMD Radeon Vega 8 Graphics
15DD, D3, AMD Radeon Vega 11 Graphics
15DD, D5, AMD Radeon Vega 8 Graphics
15DD, D6, AMD Radeon Vega 11 Graphics
15DD, D7, AMD Radeon Vega 8 Graphics
15DD, D8, AMD Radeon Vega 3 Graphics
15DD, D9, AMD Radeon Vega 6 Graphics
15DD, E1, AMD Radeon Vega 3 Graphics
15DD, E2, AMD Radeon Vega 3 Graphics
163F, AE, AMD Custom GPU 0405
163F, E1, AMD Custom GPU 0405
164E, D8, AMD Radeon 610M
164E, D9, AMD Radeon 610M
164E, DA, AMD Radeon 610M
164E, DB, AMD Radeon 610M
164E, DC, AMD Radeon 610M
1681, 06, AMD Radeon 680M
1681, 07, AMD Radeon 660M
1681, 0A, AMD Radeon 680M
1681, 0B, AMD Radeon 660M
1681, C7, AMD Radeon 680M
1681, C8, AMD Radeon 680M
1681, C9, AMD Radeon 660M
1900, 01, AMD Radeon 780M Graphics
1900, 02, AMD Radeon 760M Graphics
1900, 03, AMD Radeon 780M Graphics
1900, 04, AMD Radeon 760M Graphics
1900, 05, AMD Radeon 780M Graphics
1900, 06, AMD Radeon 780M Graphics
1900, 07, AMD Radeon 760M Graphics
1900, B0, AMD Radeon 780M Graphics
1900, B1, AMD Radeon 780M Graphics
1900, B2, AMD Radeon 780M Graphics
1900, B3, AMD Radeon 780M Graphics
1900, B4, AMD Radeon 780M Graphics
1900, B5, AMD Radeon 780M Graphics
1900, B6, AMD Radeon 780M Graphics
1900, B7, AMD Radeon 760M Graphics
1900, B8, AMD Radeon 760M Graphics
1900, B9, AMD Radeon 780M Graphics
1900, BA, AMD Radeon 780M Graphics
1900, BB, AMD Radeon 780M Graphics
1900, C0, AMD Radeon 780M Graphics
1900, C1, AMD Radeon 760M Graphics
1900, C2, AMD Radeon 780M Graphics
1900, C3, AMD Radeon 760M Graphics
1900, C4, AMD Radeon 780M Graphics
1900, C5, AMD Radeon 780M Graphics
1900, C6, AMD Radeon 760M Graphics
1900, C7, AMD Radeon 780M Graphics
1900, C8, AMD Radeon 760M Graphics
1900, C9, AMD Radeon 780M Graphics
1900, CA, AMD Radeon 760M Graphics
1900, CB, AMD Radeon 780M Graphics
1900, CC, AMD Radeon 780M Graphics
1900, CD, AMD Radeon 760M Graphics
1900, CE, AMD Radeon 780M Graphics
1900, CF, AMD Radeon 760M Graphics
1900, D0, AMD Radeon 780M Graphics
1900, D1, AMD Radeon 760M Graphics
1900, D2, AMD Radeon 780M Graphics
1900, D3, AMD Radeon 760M Graphics
1900, D4, AMD Radeon 780M Graphics
1900, D5, AMD Radeon 780M Graphics
1900, D6, AMD Radeon 760M Graphics
1900, D7, AMD Radeon 780M Graphics
1900, D8, AMD Radeon 760M Graphics
1900, D9, AMD Radeon 780M Graphics
1900, DA, AMD Radeon 760M Graphics
1900, DB, AMD Radeon 780M Graphics
1900, DC, AMD Radeon 780M Graphics
1900, DD, AMD Radeon 760M Graphics
1900, DE, AMD Radeon 780M Graphics
1900, DF, AMD Radeon 760M Graphics
1900, F0, AMD Radeon 780M Graphics
1900, F1, AMD Radeon 780M Graphics
1900, F2, AMD Radeon 780M Graphics
1901, C1, AMD Radeon 740M Graphics
1901, C2, AMD Radeon 740M Graphics
1901, C3, AMD Radeon 740M Graphics
1901, C6, AMD Radeon 740M Graphics
1901, C7, AMD Radeon 740M Graphics
1901, C8, AMD Radeon 740M Graphics
1901, C9, AMD Radeon 740M Graphics
1901, CA, AMD Radeon 740M Graphics
1901, D1, AMD Radeon 740M Graphics
1901, D2, AMD Radeon 740M Graphics
1901, D3, AMD Radeon 740M Graphics
1901, D4, AMD Radeon 740M Graphics
1901, D5, AMD Radeon 740M Graphics
1901, D6, AMD Radeon 740M Graphics
1901, D7, AMD Radeon 740M Graphics
1901, D8, AMD Radeon 740M Graphics
6600, 00, AMD Radeon HD 8600 / 8700M
6600, 81, AMD Radeon R7 M370
6601, 00, AMD Radeon HD 8500M / 8700M
6604, 00, AMD Radeon R7 M265 Series
6604, 81, AMD Radeon R7 M350
6605, 00, AMD Radeon R7 M260 Series
6605, 81, AMD Radeon R7 M340
6606, 00, AMD Radeon HD 8790M
6607, 00, AMD Radeon R5 M240
6608, 00, AMD FirePro W2100
6610, 00, AMD Radeon R7 200 Series
6610, 81, AMD Radeon R7 350
6610, 83, AMD Radeon R5 340
6610, 87, AMD Radeon R7 200 Series
6611, 00, AMD Radeon R7 200 Series
6611, 87, AMD Radeon R7 200 Series
6613, 00, AMD Radeon R7 200 Series
6617, 00, AMD Radeon R7 240 Series
6617, 87, AMD Radeon R7 200 Series
6617, C7, AMD Radeon R7 240 Series
6640, 00, AMD Radeon HD 8950
6640, 80, AMD Radeon R9 M380
6646, 00, AMD Radeon R9 M280X
6646, 80, AMD Radeon R9 M385
6646, 80, AMD Radeon R9 M470X
6647, 00, AMD Radeon R9 M200X Series
6647, 80, AMD Radeon R9 M380
6649, 00, AMD FirePro W5100
6658, 00, AMD Radeon R7 200 Series
665C, 00, AMD Radeon HD 7700 Series
665D, 00, AMD Radeon R7 200 Series
665F, 81, AMD Radeon R7 360 Series
6660, 00, AMD Radeon HD 8600M Series
6660, 81, AMD Radeon R5 M335
6660, 83, AMD Radeon R5 M330
6663, 00, AMD Radeon HD 8500M Series
6663, 83, AMD Radeon R5 M320
6664, 00, AMD Radeon R5 M200 Series
6665, 00, AMD Radeon R5 M230 Series
6665, 83, AMD Radeon R5 M320
6665, C3, AMD Radeon R5 M435
6666, 00, AMD Radeon R5 M200 Series
6667, 00, AMD Radeon R5 M200 Series
666F, 00, AMD Radeon HD 8500M
66A1, 02, AMD Instinct MI60 / MI50
66A1, 06, AMD Radeon Pro VII
66AF, C1, AMD Radeon VII
6780, 00, AMD FirePro W9000
6784, 00, ATI FirePro V (FireGL V) Graphics Adapter
6788, 00, ATI FirePro V (FireGL V) Graphics Adapter
678A, 00, AMD FirePro W8000
6798, 00, AMD Radeon R9 200 / HD 7900 Series
6799, 00, AMD Radeon HD 7900 Series
679A, 00, AMD Radeon HD 7900 Series
679B, 00, AMD Radeon HD 7900 Series
679E, 00, AMD Radeon HD 7800 Series
67A0, 00, AMD Radeon FirePro W9100
67A1, 00, AMD Radeon FirePro W8100
67B0, 00, AMD Radeon R9 200 Series
67B0, 80, AMD Radeon R9 390 Series
67B1, 00, AMD Radeon R9 200 Series
67B1, 80, AMD Radeon R9 390 Series
67B9, 00, AMD Radeon R9 200 Series
67C0, 00, AMD Radeon Pro WX 7100 Graphics
67C0, 80, AMD Radeon E9550
67C2, 01, AMD Radeon Pro V7350x2
67C2, 02, AMD Radeon Pro V7300X
67C4, 00, AMD Radeon Pro WX 7100 Graphics
67C4, 80, AMD Radeon E9560 / E9565 Graphics
67C7, 00, AMD Radeon Pro WX 5100 Graphics
67C7, 80, AMD Radeon E9390 Graphics
67D0, 01, AMD Radeon Pro V7350x2
67D0, 02, AMD Radeon Pro V7300X
67DF, C0, AMD Radeon Pro 580X
67DF, C1, AMD Radeon RX 580 Series
67DF, C2, AMD Radeon RX 570 Series
67DF, C3, AMD Radeon RX 580 Series
67DF, C4, AMD Radeon RX 480 Graphics
67DF, C5, AMD Radeon RX 470 Graphics
67DF, C6, AMD Radeon RX 570 Series
67DF, C7, AMD Radeon RX 480 Graphics
67DF, CF, AMD Radeon RX 470 Graphics
67DF, D7, AMD Radeon RX 470 Graphics
67DF, E0, AMD Radeon RX 470 Series
67DF, E1, AMD Radeon RX 590 Series
67DF, E3, AMD Radeon RX Series
67DF, E7, AMD Radeon RX 580 Series
67DF, EB, AMD Radeon Pro 580X
67DF, EF, AMD Radeon RX 570 Series
67DF, F7, AMD Radeon RX P30PH
67DF, FF, AMD Radeon RX 470 Series
67E0, 00, AMD Radeon Pro WX Series
67E3, 00, AMD Radeon Pro WX 4100
67E8, 00, AMD Radeon Pro WX Series
67E8, 01, AMD Radeon Pro WX Series
67E8, 80, AMD Radeon E9260 Graphics
67EB, 00, AMD Radeon Pro V5300X
67EF, C0, AMD Radeon RX Graphics
67EF, C1, AMD Radeon RX 460 Graphics
67EF, C2, AMD Radeon Pro Series
67EF, C3, AMD Radeon RX Series
67EF, C5, AMD Radeon RX 460 Graphics
67EF, C7, AMD Radeon RX Graphics
67EF, CF, AMD Radeon RX 460 Graphics
67EF, E0, AMD Radeon RX 560 Series
67EF, E1, AMD Radeon RX Series
67EF, E2, AMD Radeon RX 560X
67EF, E3, AMD Radeon RX Series
67EF, E5, AMD Radeon RX 560 Series
67EF, E7, AMD Radeon RX 560 Series
67EF, EF, AMD Radeon 550 Series
67EF, FF, AMD Radeon RX 460 Graphics
67FF, C0, AMD Radeon Pro 465
67FF, C1, AMD Radeon RX 560 Series
67FF, CF, AMD Radeon RX 560 Series
67FF, EF, AMD Radeon RX 560 Series
67FF, FF, AMD Radeon RX 550 Series
6800, 00, AMD Radeon HD 7970M
6801, 00, AMD Radeon HD 8970M
6806, 00, AMD Radeon R9 M290X
6808, 00, AMD FirePro W7000
6808, 00, ATI FirePro V (FireGL V) Graphics Adapter
6809, 00, ATI FirePro W5000
6810, 00, AMD Radeon R9 200 Series
6810, 81, AMD Radeon R9 370 Series
6811, 00, AMD Radeon R9 200 Series
6811, 81, AMD Radeon R7 370 Series
6818, 00, AMD Radeon HD 7800 Series
6819, 00, AMD Radeon HD 7800 Series
6820, 00, AMD Radeon R9 M275X
6820, 81, AMD Radeon R9 M375
6820, 83, AMD Radeon R9 M375X
6821, 00, AMD Radeon R9 M200X Series
6821, 83, AMD Radeon R9 M370X
6821, 87, AMD Radeon R7 M380
6822, 00, AMD Radeon E8860
6823, 00, AMD Radeon R9 M200X Series
6825, 00, AMD Radeon HD 7800M Series
6826, 00, AMD Radeon HD 7700M Series
6827, 00, AMD Radeon HD 7800M Series
6828, 00, AMD FirePro W600
682B, 00, AMD Radeon HD 8800M Series
682B, 87, AMD Radeon R9 M360
682C, 00, AMD FirePro W4100
682D, 00, AMD Radeon HD 7700M Series
682F, 00, AMD Radeon HD 7700M Series
6830, 00, AMD Radeon 7800M Series
6831, 00, AMD Radeon 7700M Series
6835, 00, AMD Radeon R7 Series / HD 9000 Series
6837, 00, AMD Radeon HD 7700 Series
683D, 00, AMD Radeon HD 7700 Series
683F, 00, AMD Radeon HD 7700 Series
684C, 00, ATI FirePro V (FireGL V) Graphics Adapter
6860, 00, AMD Radeon Instinct MI25
6860, 01, AMD Radeon Instinct MI25
6860, 02, AMD Radeon Instinct MI25
6860, 03, AMD Radeon Pro V340
6860, 04, AMD Radeon Instinct MI25x2
6860, 07, AMD Radeon Pro V320
6861, 00, AMD Radeon Pro WX 9100
6862, 00, AMD Radeon Pro SSG
6863, 00, AMD Radeon Vega Frontier Edition
6864, 03, AMD Radeon Pro V340
6864, 04, AMD Radeon Instinct MI25x2
6864, 05, AMD Radeon Pro V340
6868, 00, AMD Radeon Pro WX 8200
686C, 00, AMD Radeon Instinct MI25 MxGPU
686C, 01, AMD Radeon Instinct MI25 MxGPU
686C, 02, AMD Radeon Instinct MI25 MxGPU
686C, 03, AMD Radeon Pro V340 MxGPU
686C, 04, AMD Radeon Instinct MI25x2 MxGPU
686C, 05, AMD Radeon Pro V340L MxGPU
686C, 06, AMD Radeon Instinct MI25 MxGPU
687F, 01, AMD Radeon RX Vega
687F, C0, AMD Radeon RX Vega
687F, C1, AMD Radeon RX Vega
687F, C3, AMD Radeon RX Vega
687F, C7, AMD Radeon RX Vega
6900, 00, AMD Radeon R7 M260
6900, 81, AMD Radeon R7 M360
6900, 83, AMD Radeon R7 M340
6900, C1, AMD Radeon R5 M465 Series
6900, C3, AMD Radeon R5 M445 Series
6900, D1, AMD Radeon 530 Series
6900, D3, AMD Radeon 530 Series
6901, 00, AMD Radeon R5 M255
6902, 00, AMD Radeon Series
6907, 00, AMD Radeon R5 M255
6907, 87, AMD Radeon R5 M315
6920, 00, AMD Radeon R9 M395X
6920, 01, AMD Radeon R9 M390X
6921, 00, AMD Radeon R9 M390X
6929, 00, AMD FirePro S7150
6929, 01, AMD FirePro S7100X
692B, 00, AMD FirePro W7100
6938, 00, AMD Radeon R9 200 Series
6938, F0, AMD Radeon R9 200 Series
6938, F1, AMD Radeon R9 380 Series
6939, 00, AMD Radeon R9 200 Series
6939, F0, AMD Radeon R9 200 Series
6939, F1, AMD Radeon R9 380 Series
694C, C0, AMD Radeon RX Vega M GH Graphics
694E, C0, AMD Radeon RX Vega M GL Graphics
6980, 00, AMD Radeon Pro WX 3100
6981, 00, AMD Radeon Pro WX 3200 Series
6981, 01, AMD Radeon Pro WX 3200 Series
6981, 10, AMD Radeon Pro WX 3200 Series
6985, 00, AMD Radeon Pro WX 3100
6986, 00, AMD Radeon Pro WX 2100
6987, 80, AMD Embedded Radeon E9171
6987, C0, AMD Radeon 550X Series
6987, C1, AMD Radeon RX 640
6987, C3, AMD Radeon 540X Series
6987, C7, AMD Radeon 540
6995, 00, AMD Radeon Pro WX 2100
6997, 00, AMD Radeon Pro WX 2100
699F, 81, AMD Embedded Radeon E9170 Series
699F, C0, AMD Radeon 500 Series
699F, C1, AMD Radeon 540 Series
699F, C3, AMD Radeon 500 Series
699F, C7, AMD Radeon RX 550 / 550 Series
699F, C9, AMD Radeon 540
6FDF, E7, AMD Radeon RX 590 GME
6FDF, EF, AMD Radeon RX 580 2048SP
7300, C1, AMD FirePro S9300 x2
7300, C8, AMD Radeon R9 Fury Series
7300, C9, AMD Radeon Pro Duo
7300, CA, AMD Radeon R9 Fury Series
7300, CB, AMD Radeon R9 Fury Series
7312, 00, AMD Radeon Pro W5700
731E, C6, AMD Radeon RX 5700XTB
731E, C7, AMD Radeon RX 5700B
731F, C0, AMD Radeon RX 5700 XT 50th Anniversary
731F, C1, AMD Radeon RX 5700 XT
731F, C2, AMD Radeon RX 5600M
731F, C3, AMD Radeon RX 5700M
731F, C4, AMD Radeon RX 5700
731F, C5, AMD Radeon RX 5700 XT
731F, CA, AMD Radeon RX 5600 XT
731F, CB, AMD Radeon RX 5600 OEM
7340, C1, AMD Radeon RX 5500M
7340, C3, AMD Radeon RX 5300M
7340, C5, AMD Radeon RX 5500 XT
7340, C7, AMD Radeon RX 5500
7340, C9, AMD Radeon RX 5500XTB
7340, CF, AMD Radeon RX 5300
7341, 00, AMD Radeon Pro W5500
7347, 00, AMD Radeon Pro W5500M
7360, 41, AMD Radeon Pro 5600M
7360, C3, AMD Radeon Pro V520
7362, C1, AMD Radeon Pro V540
7362, C3, AMD Radeon Pro V520
738C, 01, AMD Instinct MI100
73A1, 00, AMD Radeon Pro V620
73A3, 00, AMD Radeon Pro W6800
73A5, C0, AMD Radeon RX 6950 XT
73AE, 00, AMD Radeon Pro V620 MxGPU
73AF, C0, AMD Radeon RX 6900 XT
73BF, C0, AMD Radeon RX 6900 XT
73BF, C1, AMD Radeon RX 6800 XT
73BF, C3, AMD Radeon RX 6800
73DF, C0, AMD Radeon RX 6750 XT
73DF, C1, AMD Radeon RX 6700 XT
73DF, C2, AMD Radeon RX 6800M
73DF, C3, AMD Radeon RX 6800M
73DF, C5, AMD Radeon RX 6700 XT
73DF, CF, AMD Radeon RX 6700M
73DF, D5, AMD Radeon RX 6750 GRE 12GB
73DF, D7, AMD TDC-235
73DF, DF, AMD Radeon RX 6700
73DF, E5, AMD Radeon RX 6750 GRE 12GB
73DF, FF, AMD Radeon RX 6700
73E0, 00, AMD Radeon RX 6600M
73E1, 00, AMD Radeon Pro W6600M
73E3, 00, AMD Radeon Pro W6600
73EF, C0, AMD Radeon RX 6800S
73EF, C1, AMD Radeon RX 6650 XT
73EF, C2, AMD Radeon RX 6700S
73EF, C3, AMD Radeon RX 6650M
73EF, C4, AMD Radeon RX 6650M XT
73FF, C1, AMD Radeon RX 6600 XT
73FF, C3, AMD Radeon RX 6600M
73FF, C7, AMD Radeon RX 6600
73FF, CB, AMD Radeon RX 6600S
73FF, CF, AMD Radeon RX 6600 LE
73FF, DF, AMD Radeon RX 6750 GRE 10GB
7408, 00, AMD Instinct MI250X
740C, 01, AMD Instinct MI250X / MI250
740F, 02, AMD Instinct MI210
7421, 00, AMD Radeon Pro W6500M
7422, 00, AMD Radeon Pro W6400
7423, 00, AMD Radeon Pro W6300M
7423, 01, AMD Radeon Pro W6300
7424, 00, AMD Radeon RX 6300
743F, C1, AMD Radeon RX 6500 XT
743F, C3, AMD Radeon RX 6500
743F, C3, AMD Radeon RX 6500M
743F, C7, AMD Radeon RX 6400
743F, C8, AMD Radeon RX 6500M
743F, CC, AMD Radeon 6550S
743F, CE, AMD Radeon RX 6450M
743F, CF, AMD Radeon RX 6300M
743F, D3, AMD Radeon RX 6550M
743F, D7, AMD Radeon RX 6400
7448, 00, AMD Radeon Pro W7900
7449, 00, AMD Radeon Pro W7800 48GB
744A, 00, AMD Radeon Pro W7900 Dual Slot
744B, 00, AMD Radeon Pro W7900D
744C, C8, AMD Radeon RX 7900 XTX
744C, CC, AMD Radeon RX 7900 XT
744C, CE, AMD Radeon RX 7900 GRE
744C, CF, AMD Radeon RX 7900M
745E, CC, AMD Radeon Pro W7800
7460, 00, AMD Radeon Pro V710
7461, 00, AMD Radeon Pro V710 MxGPU
7470, 00, AMD Radeon Pro W7700
747E, C8, AMD Radeon RX 7800 XT
747E, D8, AMD Radeon RX 7800M
747E, DB, AMD Radeon RX 7700
747E, FF, AMD Radeon RX 7700 XT
7480, 00, AMD Radeon Pro W7600
7480, C0, AMD Radeon RX 7600 XT
7480, C1, AMD Radeon RX 7700S
7480, C2, AMD Radeon RX 7650 GRE
7480, C3, AMD Radeon RX 7600S
7480, C7, AMD Radeon RX 7600M XT
7480, CF, AMD Radeon RX 7600
7481, C7, AMD Steam Machine
7483, CF, AMD Radeon RX 7600M
7489, 00, AMD Radeon Pro W7500
7499, 00, AMD Radeon Pro W7400
7499, C0, AMD Radeon RX 7400
7499, C1, AMD Radeon RX 7300
74A0, 00, AMD Instinct MI300A
74A1, 00, AMD Instinct MI300X
74A2, 00, AMD Instinct MI308X
74A5, 00, AMD Instinct MI325X
74A8, 00, AMD Instinct MI308X HF
74A9, 00, AMD Instinct MI300X HF
74B5, 00, AMD Instinct MI300X VF
74B6, 00, AMD Instinct MI308X
74BD, 00, AMD Instinct MI300X HF
7550, C0, AMD Radeon RX 9070 XT
7550, C2, AMD Radeon RX 9070 GRE
7550, C3, AMD Radeon RX 9070
7551, C0, AMD Radeon AI PRO R9700
7590, C0, AMD Radeon RX 9060 XT
7590, C7, AMD Radeon RX 9060
75A0, C0, AMD Instinct MI350X
75A3, C0, AMD Instinct MI355X
75B0, C0, AMD Instinct MI350X VF
75B3, C0, AMD Instinct MI355X VF
9830, 00, AMD Radeon HD 8400 / R3 Series
9831, 00, AMD Radeon HD 8400E
9832, 00, AMD Radeon HD 8330
9833, 00, AMD Radeon HD 8330E
9834, 00, AMD Radeon HD 8210
9835, 00, AMD Radeon HD 8210E
9836, 00, AMD Radeon HD 8200 / R3 Series
9837, 00, AMD Radeon HD 8280E
9838, 00, AMD Radeon HD 8200 / R3 series
9839, 00, AMD Radeon HD 8180
983D, 00, AMD Radeon HD 8250
9850, 00, AMD Radeon R3 Graphics
9850, 03, AMD Radeon R3 Graphics
9850, 40, AMD Radeon R2 Graphics
9850, 45, AMD Radeon R3 Graphics
9851, 00, AMD Radeon R4 Graphics
9851, 01, AMD Radeon R5E Graphics
9851, 05, AMD Radeon R5 Graphics
9851, 06, AMD Radeon R5E Graphics
9851, 40, AMD Radeon R4 Graphics
9851, 45, AMD Radeon R5 Graphics
9852, 00, AMD Radeon R2 Graphics
9852, 40, AMD Radeon E1 Graphics
9853, 00, AMD Radeon R2 Graphics
9853, 01, AMD Radeon R4E Graphics
9853, 03, AMD Radeon R2 Graphics
9853, 05, AMD Radeon R1E Graphics
9853, 06, AMD Radeon R1E Graphics
9853, 07, AMD Radeon R1E Graphics
9853, 08, AMD Radeon R1E Graphics
9853, 40, AMD Radeon R2 Graphics
9854, 00, AMD Radeon R3 Graphics
9854, 01, AMD Radeon R3E Graphics
9854, 02, AMD Radeon R3 Graphics
9854, 05, AMD Radeon R2 Graphics
9854, 06, AMD Radeon R4 Graphics
9854, 07, AMD Radeon R3 Graphics
9855, 02, AMD Radeon R6 Graphics
9855, 05, AMD Radeon R4 Graphics
9856, 00, AMD Radeon R2 Graphics
9856, 01, AMD Radeon R2E Graphics
9856, 02, AMD Radeon R2 Graphics
9856, 05, AMD Radeon R1E Graphics
9856, 06, AMD Radeon R2 Graphics
9856, 07, AMD Radeon R1E Graphics
9856, 08, AMD Radeon R1E Graphics
9856, 13, AMD Radeon R1E Graphics
9874, 81, AMD Radeon R6 Graphics
9874, 84, AMD Radeon R7 Graphics
9874, 85, AMD Radeon R6 Graphics
9874, 87, AMD Radeon R5 Graphics
9874, 88, AMD Radeon R7E Graphics
9874, 89, AMD Radeon R6E Graphics
9874, C4, AMD Radeon R7 Graphics
9874, C5, AMD Radeon R6 Graphics
9874, C6, AMD Radeon R6 Graphics
9874, C7, AMD Radeon R5 Graphics
9874, C8, AMD Radeon R7 Graphics
9874, C9, AMD Radeon R7 Graphics
9874, CA, AMD Radeon R5 Graphics
9874, CB, AMD Radeon R5 Graphics
9874, CC, AMD Radeon R7 Graphics
9874, CD, AMD Radeon R7 Graphics
9874, CE, AMD Radeon R5 Graphics
9874, E1, AMD Radeon R7 Graphics
9874, E2, AMD Radeon R7 Graphics
9874, E3, AMD Radeon R7 Graphics
9874, E4, AMD Radeon R7 Graphics
9874, E5, AMD Radeon R5 Graphics
9874, E6, AMD Radeon R5 Graphics
98E4, 80, AMD Radeon R5E Graphics
98E4, 81, AMD Radeon R4E Graphics
98E4, 83, AMD Radeon R2E Graphics
98E4, 84, AMD Radeon R2E Graphics
98E4, 86, AMD Radeon R1E Graphics
98E4, C0, AMD Radeon R4 Graphics
98E4, C1, AMD Radeon R5 Graphics
98E4, C2, AMD Radeon R4 Graphics
98E4, C4, AMD Radeon R5 Graphics
98E4, C6, AMD Radeon R5 Graphics
98E4, C8, AMD Radeon R4 Graphics
98E4, C9, AMD Radeon R4 Graphics
98E4, CA, AMD Radeon R5 Graphics
98E4, D0, AMD Radeon R2 Graphics
98E4, D1, AMD Radeon R2 Graphics
98E4, D2, AMD Radeon R2 Graphics
98E4, D4, AMD Radeon R2 Graphics
98E4, D9, AMD Radeon R5 Graphics
98E4, DA, AMD Radeon R5 Graphics
98E4, DB, AMD Radeon R3 Graphics
98E4, E1, AMD Radeon R3 Graphics
98E4, E2, AMD Radeon R3 Graphics
98E4, E9, AMD Radeon R4 Graphics
98E4, EA, AMD Radeon R4 Graphics
98E4, EB, AMD Radeon R3 Graphics
98E4, EB, AMD Radeon R4 Graphics

View File

@@ -1,34 +0,0 @@
[
{
"device_name": "NVIDIA GeForce RTX 3050 Ti Laptop GPU",
"gpu_clock": "1485MHz",
"mem_clock": "6001MHz",
"temp": "48C",
"fan_speed": null,
"power_draw": "13W",
"gpu_util": "5%",
"encode": "0%",
"decode": "0%",
"mem_util": "8%",
"mem_total": "4294967296",
"mem_used": "349372416",
"mem_free": "3945594880",
"processes" : []
},
{
"device_name": "AMD Radeon 680M",
"gpu_clock": "2200MHz",
"mem_clock": "2400MHz",
"temp": "48C",
"fan_speed": "CPU Fan",
"power_draw": "9W",
"gpu_util": "12%",
"encode": null,
"decode": "0%",
"mem_util": "7%",
"mem_total": "16929173504",
"mem_used": "1213784064",
"mem_free": "15715389440",
"processes" : []
}
]

View File

@@ -1,17 +0,0 @@
{
"ID": "7TRN:IPZB:QYBB:VPBQ:UMPP:KARE:6ZNR:XE6T:7EWV:PKF4:ZOJD:TPYS",
"Containers": 14,
"ContainersRunning": 3,
"ContainersPaused": 1,
"ContainersStopped": 10,
"Images": 508,
"Driver": "overlay2",
"KernelVersion": "6.8.0-31-generic",
"OperatingSystem": "Ubuntu 24.04 LTS",
"OSVersion": "24.04",
"OSType": "linux",
"Architecture": "x86_64",
"NCPU": 4,
"MemTotal": 2095882240,
"ServerVersion": "27.0.1"
}

View File

@@ -1,10 +1,12 @@
package agent
import (
"fmt"
"log"
"os"
"os/exec"
"runtime"
"strings"
"github.com/henrygd/beszel/internal/ghupdate"
)
@@ -63,9 +65,9 @@ func detectRestarter() restarter {
if path, err := exec.LookPath("rc-service"); err == nil {
return &openRCRestarter{cmd: path}
}
if path, err := exec.LookPath("procd"); err == nil {
return &openWRTRestarter{cmd: path}
}
if path, err := exec.LookPath("procd"); err == nil {
return &openWRTRestarter{cmd: path}
}
if path, err := exec.LookPath("service"); err == nil {
if runtime.GOOS == "freebsd" {
return &freeBSDRestarter{cmd: path}
@@ -79,7 +81,7 @@ func detectRestarter() restarter {
func Update(useMirror bool) error {
exePath, _ := os.Executable()
dataDir, err := GetDataDir()
dataDir, err := getDataDir()
if err != nil {
dataDir = os.TempDir()
}
@@ -106,12 +108,12 @@ func Update(useMirror bool) error {
}
}
// Fix SELinux context if necessary
if err := ghupdate.HandleSELinuxContext(exePath); err != nil {
// 6) Fix SELinux context if necessary
if err := handleSELinuxContext(exePath); err != nil {
ghupdate.ColorPrintf(ghupdate.ColorYellow, "Warning: SELinux context handling: %v", err)
}
// Restart service if running under a recognised init system
// 7) Restart service if running under a recognised init system
if r := detectRestarter(); r != nil {
if err := r.Restart(); err != nil {
ghupdate.ColorPrintf(ghupdate.ColorYellow, "Warning: failed to restart service: %v", err)
@@ -125,3 +127,42 @@ func Update(useMirror bool) error {
return nil
}
// handleSELinuxContext restores or applies the correct SELinux label to the binary.
func handleSELinuxContext(path string) error {
out, err := exec.Command("getenforce").Output()
if err != nil {
// SELinux not enabled or getenforce not available
return nil
}
state := strings.TrimSpace(string(out))
if state == "Disabled" {
return nil
}
ghupdate.ColorPrint(ghupdate.ColorYellow, "SELinux is enabled; applying context…")
var errs []string
// Try persistent context via semanage+restorecon
if semanagePath, err := exec.LookPath("semanage"); err == nil {
if err := exec.Command(semanagePath, "fcontext", "-a", "-t", "bin_t", path).Run(); err != nil {
errs = append(errs, "semanage fcontext failed: "+err.Error())
} else if restoreconPath, err := exec.LookPath("restorecon"); err == nil {
if err := exec.Command(restoreconPath, "-v", path).Run(); err != nil {
errs = append(errs, "restorecon failed: "+err.Error())
}
}
}
// Fallback to temporary context via chcon
if chconPath, err := exec.LookPath("chcon"); err == nil {
if err := exec.Command(chconPath, "-t", "bin_t", path).Run(); err != nil {
errs = append(errs, "chcon failed: "+err.Error())
}
}
if len(errs) > 0 {
return fmt.Errorf("SELinux context errors: %s", strings.Join(errs, "; "))
}
return nil
}

View File

@@ -6,7 +6,7 @@ import "github.com/blang/semver"
const (
// Version is the current version of the application.
Version = "0.18.4"
Version = "0.16.1"
// AppName is the name of the application.
AppName = "beszel"
)

48
go.mod
View File

@@ -1,27 +1,25 @@
module github.com/henrygd/beszel
go 1.26.0
go 1.25.3
require (
github.com/blang/semver v3.5.1+incompatible
github.com/coreos/go-systemd/v22 v22.7.0
github.com/coreos/go-systemd/v22 v22.6.0
github.com/distatus/battery v0.11.0
github.com/ebitengine/purego v0.9.1
github.com/fxamacker/cbor/v2 v2.9.0
github.com/gliderlabs/ssh v0.3.8
github.com/google/uuid v1.6.0
github.com/lxzan/gws v1.8.9
github.com/nicholas-fedor/shoutrrr v0.13.2
github.com/pocketbase/dbx v1.12.0
github.com/pocketbase/pocketbase v0.36.4
github.com/shirou/gopsutil/v4 v4.26.1
github.com/nicholas-fedor/shoutrrr v0.12.0
github.com/pocketbase/dbx v1.11.0
github.com/pocketbase/pocketbase v0.33.0
github.com/shirou/gopsutil/v4 v4.25.10
github.com/spf13/cast v1.10.0
github.com/spf13/cobra v1.10.2
github.com/spf13/cobra v1.10.1
github.com/spf13/pflag v1.0.10
github.com/stretchr/testify v1.11.1
golang.org/x/crypto v0.48.0
golang.org/x/exp v0.0.0-20260218203240-3dfff04db8fa
golang.org/x/sys v0.41.0
golang.org/x/crypto v0.44.0
golang.org/x/exp v0.0.0-20251113190631-e25ba8c21ef6
gopkg.in/yaml.v3 v3.0.1
)
@@ -33,17 +31,18 @@ require (
github.com/dolthub/maphash v0.1.0 // indirect
github.com/domodwyer/mailyak/v3 v3.6.2 // indirect
github.com/dustin/go-humanize v1.0.1 // indirect
github.com/ebitengine/purego v0.9.1 // indirect
github.com/fatih/color v1.18.0 // indirect
github.com/gabriel-vasile/mimetype v1.4.13 // indirect
github.com/gabriel-vasile/mimetype v1.4.11 // indirect
github.com/ganigeorgiev/fexpr v0.5.0 // indirect
github.com/go-ole/go-ole v1.3.0 // indirect
github.com/go-ozzo/ozzo-validation/v4 v4.3.0 // indirect
github.com/go-sql-driver/mysql v1.9.1 // indirect
github.com/godbus/dbus/v5 v5.2.2 // indirect
github.com/golang-jwt/jwt/v5 v5.3.1 // indirect
github.com/godbus/dbus/v5 v5.2.0 // indirect
github.com/golang-jwt/jwt/v5 v5.3.0 // indirect
github.com/inconshreveable/mousetrap v1.1.0 // indirect
github.com/klauspost/compress v1.18.4 // indirect
github.com/lufia/plan9stats v0.0.0-20260216142805-b3301c5f2a88 // indirect
github.com/klauspost/compress v1.18.1 // indirect
github.com/lufia/plan9stats v0.0.0-20251013123823-9fd1530e3ec3 // indirect
github.com/mattn/go-colorable v0.1.14 // indirect
github.com/mattn/go-isatty v0.0.20 // indirect
github.com/ncruces/go-strftime v1.0.0 // indirect
@@ -54,15 +53,16 @@ require (
github.com/tklauser/numcpus v0.11.0 // indirect
github.com/x448/float16 v0.8.4 // indirect
github.com/yusufpapurcu/wmi v1.2.4 // indirect
golang.org/x/image v0.36.0 // indirect
golang.org/x/net v0.50.0 // indirect
golang.org/x/oauth2 v0.35.0 // indirect
golang.org/x/sync v0.19.0 // indirect
golang.org/x/term v0.40.0 // indirect
golang.org/x/text v0.34.0 // indirect
golang.org/x/image v0.33.0 // indirect
golang.org/x/net v0.47.0 // indirect
golang.org/x/oauth2 v0.33.0 // indirect
golang.org/x/sync v0.18.0 // indirect
golang.org/x/sys v0.38.0 // indirect
golang.org/x/term v0.37.0 // indirect
golang.org/x/text v0.31.0 // indirect
howett.net/plist v1.0.1 // indirect
modernc.org/libc v1.67.6 // indirect
modernc.org/libc v1.66.10 // indirect
modernc.org/mathutil v1.7.1 // indirect
modernc.org/memory v1.11.0 // indirect
modernc.org/sqlite v1.45.0 // indirect
modernc.org/sqlite v1.40.0 // indirect
)

114
go.sum
View File

@@ -9,8 +9,8 @@ github.com/asaskevich/govalidator v0.0.0-20230301143203-a9d515a09cc2 h1:DklsrG3d
github.com/asaskevich/govalidator v0.0.0-20230301143203-a9d515a09cc2/go.mod h1:WaHUgvxTVq04UNunO+XhnAqY/wQc+bxr74GqbsZ/Jqw=
github.com/blang/semver v3.5.1+incompatible h1:cQNTCjp13qL8KC3Nbxr/y2Bqb63oX6wdnnjpJbkM4JQ=
github.com/blang/semver v3.5.1+incompatible/go.mod h1:kRBLl5iJ+tD4TcOOxsy/0fnwebNt5EWlYSAyrTnjyyk=
github.com/coreos/go-systemd/v22 v22.7.0 h1:LAEzFkke61DFROc7zNLX/WA2i5J8gYqe0rSj9KI28KA=
github.com/coreos/go-systemd/v22 v22.7.0/go.mod h1:xNUYtjHu2EDXbsxz1i41wouACIwT7Ybq9o0BQhMwD0w=
github.com/coreos/go-systemd/v22 v22.6.0 h1:aGVa/v8B7hpb0TKl0MWoAavPDmHvobFe5R5zn0bCJWo=
github.com/coreos/go-systemd/v22 v22.6.0/go.mod h1:iG+pp635Fo7ZmV/j14KUcmEyWF+0X7Lua8rrTWzYgWU=
github.com/cpuguy83/go-md2man/v2 v2.0.6/go.mod h1:oOW0eioCTA6cOiMLiUPZOpcVxMig6NIQQ7OS05n1F4g=
github.com/davecgh/go-spew v1.1.0/go.mod h1:J7Y8YcW2NihsgmVo/mv3lAwl/skON4iLHjSsI+c5H38=
github.com/davecgh/go-spew v1.1.2-0.20180830191138-d8f796af33cc h1:U9qPSI2PIWSS1VwoXQT9A3Wy9MM3WgvqSxFWenqJduM=
@@ -33,8 +33,8 @@ github.com/frankban/quicktest v1.14.6 h1:7Xjx+VpznH+oBnejlPUj8oUpdxnVs4f8XU8WnHk
github.com/frankban/quicktest v1.14.6/go.mod h1:4ptaffx2x8+WTWXmUCuVU6aPUX1/Mz7zb5vbUoiM6w0=
github.com/fxamacker/cbor/v2 v2.9.0 h1:NpKPmjDBgUfBms6tr6JZkTHtfFGcMKsw3eGcmD/sapM=
github.com/fxamacker/cbor/v2 v2.9.0/go.mod h1:vM4b+DJCtHn+zz7h3FFp/hDAI9WNWCsZj23V5ytsSxQ=
github.com/gabriel-vasile/mimetype v1.4.13 h1:46nXokslUBsAJE/wMsp5gtO500a4F3Nkz9Ufpk2AcUM=
github.com/gabriel-vasile/mimetype v1.4.13/go.mod h1:d+9Oxyo1wTzWdyVUPMmXFvp4F9tea18J8ufA774AB3s=
github.com/gabriel-vasile/mimetype v1.4.11 h1:AQvxbp830wPhHTqc1u7nzoLT+ZFxGY7emj5DR5DYFik=
github.com/gabriel-vasile/mimetype v1.4.11/go.mod h1:d+9Oxyo1wTzWdyVUPMmXFvp4F9tea18J8ufA774AB3s=
github.com/ganigeorgiev/fexpr v0.5.0 h1:XA9JxtTE/Xm+g/JFI6RfZEHSiQlk+1glLvRK1Lpv/Tk=
github.com/ganigeorgiev/fexpr v0.5.0/go.mod h1:RyGiGqmeXhEQ6+mlGdnUleLHgtzzu/VGO2WtJkF5drE=
github.com/gliderlabs/ssh v0.3.8 h1:a4YXD1V7xMF9g5nTkdfnja3Sxy1PVDCj1Zg4Wb8vY6c=
@@ -51,15 +51,15 @@ github.com/go-sql-driver/mysql v1.9.1 h1:FrjNGn/BsJQjVRuSa8CBrM5BWA9BWoXXat3KrtS
github.com/go-sql-driver/mysql v1.9.1/go.mod h1:qn46aNg1333BRMNU69Lq93t8du/dwxI64Gl8i5p1WMU=
github.com/go-task/slim-sprig/v3 v3.0.0 h1:sUs3vkvUymDpBKi3qH1YSqBQk9+9D/8M2mN1vB6EwHI=
github.com/go-task/slim-sprig/v3 v3.0.0/go.mod h1:W848ghGpv3Qj3dhTPRyJypKRiqCdHZiAzKg9hl15HA8=
github.com/godbus/dbus/v5 v5.2.2 h1:TUR3TgtSVDmjiXOgAAyaZbYmIeP3DPkld3jgKGV8mXQ=
github.com/godbus/dbus/v5 v5.2.2/go.mod h1:3AAv2+hPq5rdnr5txxxRwiGjPXamgoIHgz9FPBfOp3c=
github.com/golang-jwt/jwt/v5 v5.3.1 h1:kYf81DTWFe7t+1VvL7eS+jKFVWaUnK9cB1qbwn63YCY=
github.com/golang-jwt/jwt/v5 v5.3.1/go.mod h1:fxCRLWMO43lRc8nhHWY6LGqRcf+1gQWArsqaEUEa5bE=
github.com/godbus/dbus/v5 v5.2.0 h1:3WexO+U+yg9T70v9FdHr9kCxYlazaAXUhx2VMkbfax8=
github.com/godbus/dbus/v5 v5.2.0/go.mod h1:3AAv2+hPq5rdnr5txxxRwiGjPXamgoIHgz9FPBfOp3c=
github.com/golang-jwt/jwt/v5 v5.3.0 h1:pv4AsKCKKZuqlgs5sUmn4x8UlGa0kEVt/puTpKx9vvo=
github.com/golang-jwt/jwt/v5 v5.3.0/go.mod h1:fxCRLWMO43lRc8nhHWY6LGqRcf+1gQWArsqaEUEa5bE=
github.com/golang/protobuf v1.3.1/go.mod h1:6lQm79b+lXiMfvg/cZm0SGofjICqVBUtrP5yJMmIC1U=
github.com/google/go-cmp v0.7.0 h1:wk8382ETsv4JYUZwIsn6YpYiWiBsYLSJiTsyBybVuN8=
github.com/google/go-cmp v0.7.0/go.mod h1:pXiqmnSA92OHEEa9HXL2W4E7lf9JzCmGVUdgjX3N/iU=
github.com/google/pprof v0.0.0-20260115054156-294ebfa9ad83 h1:z2ogiKUYzX5Is6zr/vP9vJGqPwcdqsWjOt+V8J7+bTc=
github.com/google/pprof v0.0.0-20260115054156-294ebfa9ad83/go.mod h1:MxpfABSjhmINe3F1It9d+8exIHFvUqtLIRCdOGNXqiI=
github.com/google/pprof v0.0.0-20251007162407-5df77e3f7d1d h1:KJIErDwbSHjnp/SGzE5ed8Aol7JsKiI5X7yWKAtzhM0=
github.com/google/pprof v0.0.0-20251007162407-5df77e3f7d1d/go.mod h1:I6V7YzU0XDpsHqbsyrghnFZLO1gwK6NPTNvmetQIk9U=
github.com/google/uuid v1.6.0 h1:NIvaJDMOsjHA8n1jAhLSgzrAzy1Hgr+hNrb57e+94F0=
github.com/google/uuid v1.6.0/go.mod h1:TIyPZe4MgqvfeYDBFedMoGGpEw/LqOeaOT+nhxU+yHo=
github.com/hashicorp/golang-lru/v2 v2.0.7 h1:a+bsQ5rvGLjzHuww6tVxozPZFVghXaHOwFs4luLUK2k=
@@ -69,14 +69,14 @@ github.com/inconshreveable/mousetrap v1.1.0/go.mod h1:vpF70FUmC8bwa3OWnCshd2FqLf
github.com/jarcoal/httpmock v1.4.1 h1:0Ju+VCFuARfFlhVXFc2HxlcQkfB+Xq12/EotHko+x2A=
github.com/jarcoal/httpmock v1.4.1/go.mod h1:ftW1xULwo+j0R0JJkJIIi7UKigZUXCLLanykgjwBXL0=
github.com/jessevdk/go-flags v1.4.0/go.mod h1:4FA24M0QyGHXBuZZK/XkWh8h0e1EYbRYJSGM75WSRxI=
github.com/klauspost/compress v1.18.4 h1:RPhnKRAQ4Fh8zU2FY/6ZFDwTVTxgJ/EMydqSTzE9a2c=
github.com/klauspost/compress v1.18.4/go.mod h1:R0h/fSBs8DE4ENlcrlib3PsXS61voFxhIs2DeRhCvJ4=
github.com/klauspost/compress v1.18.1 h1:bcSGx7UbpBqMChDtsF28Lw6v/G94LPrrbMbdC3JH2co=
github.com/klauspost/compress v1.18.1/go.mod h1:ZQFFVG+MdnR0P+l6wpXgIL4NTtwiKIdBnrBd8Nrxr+0=
github.com/kr/pretty v0.3.1 h1:flRD4NNwYAUpkphVc1HcthR4KEIFJ65n8Mw5qdRn3LE=
github.com/kr/pretty v0.3.1/go.mod h1:hoEshYVHaxMs3cyo3Yncou5ZscifuDolrwPKZanG3xk=
github.com/kr/text v0.2.0 h1:5Nx0Ya0ZqY2ygV366QzturHI13Jq95ApcVaJBhpS+AY=
github.com/kr/text v0.2.0/go.mod h1:eLer722TekiGuMkidMxC/pM04lWEeraHUUmBw8l2grE=
github.com/lufia/plan9stats v0.0.0-20260216142805-b3301c5f2a88 h1:PTw+yKnXcOFCR6+8hHTyWBeQ/P4Nb7dd4/0ohEcWQuM=
github.com/lufia/plan9stats v0.0.0-20260216142805-b3301c5f2a88/go.mod h1:autxFIvghDt3jPTLoqZ9OZ7s9qTGNAWmYCjVFWPX/zg=
github.com/lufia/plan9stats v0.0.0-20251013123823-9fd1530e3ec3 h1:PwQumkgq4/acIiZhtifTV5OUqqiP82UAl0h87xj/l9k=
github.com/lufia/plan9stats v0.0.0-20251013123823-9fd1530e3ec3/go.mod h1:autxFIvghDt3jPTLoqZ9OZ7s9qTGNAWmYCjVFWPX/zg=
github.com/lxzan/gws v1.8.9 h1:VU3SGUeWlQrEwfUSfokcZep8mdg/BrUF+y73YYshdBM=
github.com/lxzan/gws v1.8.9/go.mod h1:d9yHaR1eDTBHagQC6KY7ycUOaz5KWeqQtP3xu7aMK8Y=
github.com/mattn/go-colorable v0.1.14 h1:9A9LHSqF/7dyVVX6g0U9cwm9pG3kP9gSzcuIPHPsaIE=
@@ -85,19 +85,19 @@ github.com/mattn/go-isatty v0.0.20 h1:xfD0iDuEKnDkl03q4limB+vH+GxLEtL/jb4xVJSWWE
github.com/mattn/go-isatty v0.0.20/go.mod h1:W+V8PltTTMOvKvAeJH7IuucS94S2C6jfK/D7dTCTo3Y=
github.com/ncruces/go-strftime v1.0.0 h1:HMFp8mLCTPp341M/ZnA4qaf7ZlsbTc+miZjCLOFAw7w=
github.com/ncruces/go-strftime v1.0.0/go.mod h1:Fwc5htZGVVkseilnfgOVb9mKy6w1naJmn9CehxcKcls=
github.com/nicholas-fedor/shoutrrr v0.13.2 h1:hfsYBIqSFYGg92pZP5CXk/g7/OJIkLYmiUnRl+AD1IA=
github.com/nicholas-fedor/shoutrrr v0.13.2/go.mod h1:ZqzV3gY/Wj6AvWs1etlO7+yKbh4iptSbeL8avBpMQbA=
github.com/onsi/ginkgo/v2 v2.28.1 h1:S4hj+HbZp40fNKuLUQOYLDgZLwNUVn19N3Atb98NCyI=
github.com/onsi/ginkgo/v2 v2.28.1/go.mod h1:CLtbVInNckU3/+gC8LzkGUb9oF+e8W8TdUsxPwvdOgE=
github.com/onsi/gomega v1.39.1 h1:1IJLAad4zjPn2PsnhH70V4DKRFlrCzGBNrNaru+Vf28=
github.com/onsi/gomega v1.39.1/go.mod h1:hL6yVALoTOxeWudERyfppUcZXjMwIMLnuSfruD2lcfg=
github.com/nicholas-fedor/shoutrrr v0.12.0 h1:8mwJdfU+uBEybSymwQJMGl/grG7lvVUKbVSNxn3XvUI=
github.com/nicholas-fedor/shoutrrr v0.12.0/go.mod h1:WYiRalR4C43Qmd2zhPWGIFIxu633NB1hDM6Ap/DQcsA=
github.com/onsi/ginkgo/v2 v2.27.2 h1:LzwLj0b89qtIy6SSASkzlNvX6WktqurSHwkk2ipF/Ns=
github.com/onsi/ginkgo/v2 v2.27.2/go.mod h1:ArE1D/XhNXBXCBkKOLkbsb2c81dQHCRcF5zwn/ykDRo=
github.com/onsi/gomega v1.38.2 h1:eZCjf2xjZAqe+LeWvKb5weQ+NcPwX84kqJ0cZNxok2A=
github.com/onsi/gomega v1.38.2/go.mod h1:W2MJcYxRGV63b418Ai34Ud0hEdTVXq9NW9+Sx6uXf3k=
github.com/pmezard/go-difflib v1.0.0/go.mod h1:iKH77koFhYxTK1pcRnkKkqfTogsbg7gZNVY4sRDYZ/4=
github.com/pmezard/go-difflib v1.0.1-0.20181226105442-5d4384ee4fb2 h1:Jamvg5psRIccs7FGNTlIRMkT8wgtp5eCXdBlqhYGL6U=
github.com/pmezard/go-difflib v1.0.1-0.20181226105442-5d4384ee4fb2/go.mod h1:iKH77koFhYxTK1pcRnkKkqfTogsbg7gZNVY4sRDYZ/4=
github.com/pocketbase/dbx v1.12.0 h1:/oLErM+A0b4xI0PWTGPqSDVjzix48PqI/bng2l0PzoA=
github.com/pocketbase/dbx v1.12.0/go.mod h1:xXRCIAKTHMgUCyCKZm55pUOdvFziJjQfXaWKhu2vhMs=
github.com/pocketbase/pocketbase v0.36.4 h1:zTjRZbp2WfTOJJfb+pFRWa200UaQwxZYt8RzkFMlAZ4=
github.com/pocketbase/pocketbase v0.36.4/go.mod h1:9CiezhRudd9FZGa5xZa53QZBTNxc5vvw/FGG+diAECI=
github.com/pocketbase/dbx v1.11.0 h1:LpZezioMfT3K4tLrqA55wWFw1EtH1pM4tzSVa7kgszU=
github.com/pocketbase/dbx v1.11.0/go.mod h1:xXRCIAKTHMgUCyCKZm55pUOdvFziJjQfXaWKhu2vhMs=
github.com/pocketbase/pocketbase v0.33.0 h1:v2EfiY3hxigzRJ/BwFuwVn0vUv7d2QQoD5zUFPaKR9o=
github.com/pocketbase/pocketbase v0.33.0/go.mod h1:9BEs+CRV7CrS+X5LfBh4bdJQsbzQAIklft3ovGe/c5A=
github.com/power-devops/perfstat v0.0.0-20240221224432-82ca36839d55 h1:o4JXh1EVt9k/+g42oCprj/FisM4qX9L3sZB3upGN2ZU=
github.com/power-devops/perfstat v0.0.0-20240221224432-82ca36839d55/go.mod h1:OmDBASR4679mdNQnz2pUhc2G8CO2JrUAVFDRBDP/hJE=
github.com/remyoudompheng/bigfft v0.0.0-20230129092748-24d4a6f8daec h1:W09IVJc94icq4NjY3clb7Lk8O1qJ8BdBEF8z0ibU0rE=
@@ -105,12 +105,12 @@ github.com/remyoudompheng/bigfft v0.0.0-20230129092748-24d4a6f8daec/go.mod h1:qq
github.com/rogpeppe/go-internal v1.9.0 h1:73kH8U+JUqXU8lRuOHeVHaa/SZPifC7BkcraZVejAe8=
github.com/rogpeppe/go-internal v1.9.0/go.mod h1:WtVeX8xhTBvf0smdhujwtBcq4Qrzq/fJaraNFVN+nFs=
github.com/russross/blackfriday/v2 v2.1.0/go.mod h1:+Rmxgy9KzJVeS9/2gXHxylqXiyQDYRxCVz55jmeOWTM=
github.com/shirou/gopsutil/v4 v4.26.1 h1:TOkEyriIXk2HX9d4isZJtbjXbEjf5qyKPAzbzY0JWSo=
github.com/shirou/gopsutil/v4 v4.26.1/go.mod h1:medLI9/UNAb0dOI9Q3/7yWSqKkj00u+1tgY8nvv41pc=
github.com/shirou/gopsutil/v4 v4.25.10 h1:at8lk/5T1OgtuCp+AwrDofFRjnvosn0nkN2OLQ6g8tA=
github.com/shirou/gopsutil/v4 v4.25.10/go.mod h1:+kSwyC8DRUD9XXEHCAFjK+0nuArFJM0lva+StQAcskM=
github.com/spf13/cast v1.10.0 h1:h2x0u2shc1QuLHfxi+cTJvs30+ZAHOGRic8uyGTDWxY=
github.com/spf13/cast v1.10.0/go.mod h1:jNfB8QC9IA6ZuY2ZjDp0KtFO2LZZlg4S/7bzP6qqeHo=
github.com/spf13/cobra v1.10.2 h1:DMTTonx5m65Ic0GOoRY2c16WCbHxOOw6xxezuLaBpcU=
github.com/spf13/cobra v1.10.2/go.mod h1:7C1pvHqHw5A4vrJfjNwvOdzYu0Gml16OCs2GRiTUUS4=
github.com/spf13/cobra v1.10.1 h1:lJeBwCfmrnXthfAupyUTzJ/J4Nc1RsHC/mSRU2dll/s=
github.com/spf13/cobra v1.10.1/go.mod h1:7SmJGaTHFVBY0jW4NXGluQoLvhqFQM+6XSKD+P4XaB0=
github.com/spf13/pflag v1.0.9/go.mod h1:McXfInJRrz4CZXVZOBLb0bTZqETkiAhM9Iw0y3An2Bg=
github.com/spf13/pflag v1.0.10 h1:4EBh2KAYBwaONj6b2Ye1GiHfwjqyROoF4RwYO+vPwFk=
github.com/spf13/pflag v1.0.10/go.mod h1:McXfInJRrz4CZXVZOBLb0bTZqETkiAhM9Iw0y3An2Bg=
@@ -129,41 +129,41 @@ github.com/yusufpapurcu/wmi v1.2.4/go.mod h1:SBZ9tNy3G9/m5Oi98Zks0QjeHVDvuK0qfxQ
go.yaml.in/yaml/v3 v3.0.4 h1:tfq32ie2Jv2UxXFdLJdh3jXuOzWiL1fo0bu/FbuKpbc=
go.yaml.in/yaml/v3 v3.0.4/go.mod h1:DhzuOOF2ATzADvBadXxruRBLzYTpT36CKvDb3+aBEFg=
golang.org/x/crypto v0.0.0-20190308221718-c2843e01d9a2/go.mod h1:djNgcEr1/C05ACkg1iLfiJU5Ep61QUkGW8qpdssI0+w=
golang.org/x/crypto v0.48.0 h1:/VRzVqiRSggnhY7gNRxPauEQ5Drw9haKdM0jqfcCFts=
golang.org/x/crypto v0.48.0/go.mod h1:r0kV5h3qnFPlQnBSrULhlsRfryS2pmewsg+XfMgkVos=
golang.org/x/exp v0.0.0-20260218203240-3dfff04db8fa h1:Zt3DZoOFFYkKhDT3v7Lm9FDMEV06GpzjG2jrqW+QTE0=
golang.org/x/exp v0.0.0-20260218203240-3dfff04db8fa/go.mod h1:K79w1Vqn7PoiZn+TkNpx3BUWUQksGO3JcVX6qIjytmA=
golang.org/x/crypto v0.44.0 h1:A97SsFvM3AIwEEmTBiaxPPTYpDC47w720rdiiUvgoAU=
golang.org/x/crypto v0.44.0/go.mod h1:013i+Nw79BMiQiMsOPcVCB5ZIJbYkerPrGnOa00tvmc=
golang.org/x/exp v0.0.0-20251113190631-e25ba8c21ef6 h1:zfMcR1Cs4KNuomFFgGefv5N0czO2XZpUbxGUy8i8ug0=
golang.org/x/exp v0.0.0-20251113190631-e25ba8c21ef6/go.mod h1:46edojNIoXTNOhySWIWdix628clX9ODXwPsQuG6hsK0=
golang.org/x/image v0.0.0-20191009234506-e7c1f5e7dbb8/go.mod h1:FeLwcggjj3mMvU+oOTbSwawSJRM1uh48EjtB4UJZlP0=
golang.org/x/image v0.36.0 h1:Iknbfm1afbgtwPTmHnS2gTM/6PPZfH+z2EFuOkSbqwc=
golang.org/x/image v0.36.0/go.mod h1:YsWD2TyyGKiIX1kZlu9QfKIsQ4nAAK9bdgdrIsE7xy4=
golang.org/x/mod v0.33.0 h1:tHFzIWbBifEmbwtGz65eaWyGiGZatSrT9prnU8DbVL8=
golang.org/x/mod v0.33.0/go.mod h1:swjeQEj+6r7fODbD2cqrnje9PnziFuw4bmLbBZFrQ5w=
golang.org/x/image v0.33.0 h1:LXRZRnv1+zGd5XBUVRFmYEphyyKJjQjCRiOuAP3sZfQ=
golang.org/x/image v0.33.0/go.mod h1:DD3OsTYT9chzuzTQt+zMcOlBHgfoKQb1gry8p76Y1sc=
golang.org/x/mod v0.30.0 h1:fDEXFVZ/fmCKProc/yAXXUijritrDzahmwwefnjoPFk=
golang.org/x/mod v0.30.0/go.mod h1:lAsf5O2EvJeSFMiBxXDki7sCgAxEUcZHXoXMKT4GJKc=
golang.org/x/net v0.0.0-20190603091049-60506f45cf65/go.mod h1:HSz+uSET+XFnRR8LxR5pz3Of3rY3CfYBVs4xY44aLks=
golang.org/x/net v0.50.0 h1:ucWh9eiCGyDR3vtzso0WMQinm2Dnt8cFMuQa9K33J60=
golang.org/x/net v0.50.0/go.mod h1:UgoSli3F/pBgdJBHCTc+tp3gmrU4XswgGRgtnwWTfyM=
golang.org/x/oauth2 v0.35.0 h1:Mv2mzuHuZuY2+bkyWXIHMfhNdJAdwW3FuWeCPYN5GVQ=
golang.org/x/oauth2 v0.35.0/go.mod h1:lzm5WQJQwKZ3nwavOZ3IS5Aulzxi68dUSgRHujetwEA=
golang.org/x/sync v0.19.0 h1:vV+1eWNmZ5geRlYjzm2adRgW2/mcpevXNg50YZtPCE4=
golang.org/x/sync v0.19.0/go.mod h1:9KTHXmSnoGruLpwFjVSX0lNNA75CykiMECbovNTZqGI=
golang.org/x/net v0.47.0 h1:Mx+4dIFzqraBXUugkia1OOvlD6LemFo1ALMHjrXDOhY=
golang.org/x/net v0.47.0/go.mod h1:/jNxtkgq5yWUGYkaZGqo27cfGZ1c5Nen03aYrrKpVRU=
golang.org/x/oauth2 v0.33.0 h1:4Q+qn+E5z8gPRJfmRy7C2gGG3T4jIprK6aSYgTXGRpo=
golang.org/x/oauth2 v0.33.0/go.mod h1:lzm5WQJQwKZ3nwavOZ3IS5Aulzxi68dUSgRHujetwEA=
golang.org/x/sync v0.18.0 h1:kr88TuHDroi+UVf+0hZnirlk8o8T+4MrK6mr60WkH/I=
golang.org/x/sync v0.18.0/go.mod h1:9KTHXmSnoGruLpwFjVSX0lNNA75CykiMECbovNTZqGI=
golang.org/x/sys v0.0.0-20190215142949-d0b11bdaac8a/go.mod h1:STP8DvDyc/dI5b8T5hshtkjS+E42TnysNCUPdjciGhY=
golang.org/x/sys v0.0.0-20190916202348-b4ddaad3f8a3/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
golang.org/x/sys v0.0.0-20201204225414-ed752295db88/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
golang.org/x/sys v0.1.0/go.mod h1:oPkhp1MJrh7nUepCBck5+mAzfO9JrbApNNgaTdGDITg=
golang.org/x/sys v0.6.0/go.mod h1:oPkhp1MJrh7nUepCBck5+mAzfO9JrbApNNgaTdGDITg=
golang.org/x/sys v0.41.0 h1:Ivj+2Cp/ylzLiEU89QhWblYnOE9zerudt9Ftecq2C6k=
golang.org/x/sys v0.41.0/go.mod h1:OgkHotnGiDImocRcuBABYBEXf8A9a87e/uXjp9XT3ks=
golang.org/x/term v0.40.0 h1:36e4zGLqU4yhjlmxEaagx2KuYbJq3EwY8K943ZsHcvg=
golang.org/x/term v0.40.0/go.mod h1:w2P8uVp06p2iyKKuvXIm7N/y0UCRt3UfJTfZ7oOpglM=
golang.org/x/sys v0.38.0 h1:3yZWxaJjBmCWXqhN1qh02AkOnCQ1poK6oF+a7xWL6Gc=
golang.org/x/sys v0.38.0/go.mod h1:OgkHotnGiDImocRcuBABYBEXf8A9a87e/uXjp9XT3ks=
golang.org/x/term v0.37.0 h1:8EGAD0qCmHYZg6J17DvsMy9/wJ7/D/4pV/wfnld5lTU=
golang.org/x/term v0.37.0/go.mod h1:5pB4lxRNYYVZuTLmy8oR2BH8dflOR+IbTYFD8fi3254=
golang.org/x/text v0.3.0/go.mod h1:NqM8EUOU14njkJ3fqMW+pc6Ldnwhi/IjpwHt7yyuwOQ=
golang.org/x/text v0.3.2/go.mod h1:bEr9sfX3Q8Zfm5fL9x+3itogRgK3+ptLWKqgva+5dAk=
golang.org/x/text v0.34.0 h1:oL/Qq0Kdaqxa1KbNeMKwQq0reLCCaFtqu2eNuSeNHbk=
golang.org/x/text v0.34.0/go.mod h1:homfLqTYRFyVYemLBFl5GgL/DWEiH5wcsQ5gSh1yziA=
golang.org/x/text v0.31.0 h1:aC8ghyu4JhP8VojJ2lEHBnochRno1sgL6nEi9WGFGMM=
golang.org/x/text v0.31.0/go.mod h1:tKRAlv61yKIjGGHX/4tP1LTbc13YSec1pxVEWXzfoeM=
golang.org/x/tools v0.0.0-20180917221912-90fa682c2a6e/go.mod h1:n7NCudcB/nEzxVGmLbDWY5pfWTLqBcC2KZ6jyYvM4mQ=
golang.org/x/tools v0.42.0 h1:uNgphsn75Tdz5Ji2q36v/nsFSfR/9BRFvqhGBaJGd5k=
golang.org/x/tools v0.42.0/go.mod h1:Ma6lCIwGZvHK6XtgbswSoWroEkhugApmsXyrUmBhfr0=
golang.org/x/tools v0.39.0 h1:ik4ho21kwuQln40uelmciQPp9SipgNDdrafrYA4TmQQ=
golang.org/x/tools v0.39.0/go.mod h1:JnefbkDPyD8UU2kI5fuf8ZX4/yUeh9W877ZeBONxUqQ=
google.golang.org/appengine v1.6.5/go.mod h1:8WjMMxjGQR8xUklV/ARdw2HLXBOI7O7uCIDZVag1xfc=
google.golang.org/protobuf v1.36.11 h1:fV6ZwhNocDyBLK0dj+fg8ektcVegBBuEolpbTQyBNVE=
google.golang.org/protobuf v1.36.11/go.mod h1:HTf+CrKn2C3g5S8VImy6tdcUvCska2kB7j23XfzDpco=
google.golang.org/protobuf v1.36.10 h1:AYd7cD/uASjIL6Q9LiTjz8JLcrh/88q5UObnmY3aOOE=
google.golang.org/protobuf v1.36.10/go.mod h1:HTf+CrKn2C3g5S8VImy6tdcUvCska2kB7j23XfzDpco=
gopkg.in/check.v1 v0.0.0-20161208181325-20d25e280405/go.mod h1:Co6ibVJAznAaIkqp8huTwlJQCZ016jof/cbN4VW5Yz0=
gopkg.in/check.v1 v1.0.0-20201130134442-10cb98267c6c h1:Hei/4ADfdWqJk1ZMxUNpqntNwaWcugrBjAiHlqqRiVk=
gopkg.in/check.v1 v1.0.0-20201130134442-10cb98267c6c/go.mod h1:JHkPIbrfpd72SG/EVd6muEfDQjcINNoR0C8j2r3qZ4Q=
@@ -185,8 +185,10 @@ modernc.org/gc/v3 v3.1.1 h1:k8T3gkXWY9sEiytKhcgyiZ2L0DTyCQ/nvX+LoCljoRE=
modernc.org/gc/v3 v3.1.1/go.mod h1:HFK/6AGESC7Ex+EZJhJ2Gni6cTaYpSMmU/cT9RmlfYY=
modernc.org/goabi0 v0.2.0 h1:HvEowk7LxcPd0eq6mVOAEMai46V+i7Jrj13t4AzuNks=
modernc.org/goabi0 v0.2.0/go.mod h1:CEFRnnJhKvWT1c1JTI3Avm+tgOWbkOu5oPA8eH8LnMI=
modernc.org/libc v1.67.6 h1:eVOQvpModVLKOdT+LvBPjdQqfrZq+pC39BygcT+E7OI=
modernc.org/libc v1.67.6/go.mod h1:JAhxUVlolfYDErnwiqaLvUqc8nfb2r6S6slAgZOnaiE=
modernc.org/libc v1.66.10 h1:yZkb3YeLx4oynyR+iUsXsybsX4Ubx7MQlSYEw4yj59A=
modernc.org/libc v1.66.10/go.mod h1:8vGSEwvoUoltr4dlywvHqjtAqHBaw0j1jI7iFBTAr2I=
modernc.org/libc v1.67.0 h1:QzL4IrKab2OFmxA3/vRYl0tLXrIamwrhD6CKD4WBVjQ=
modernc.org/libc v1.67.0/go.mod h1:QvvnnJ5P7aitu0ReNpVIEyesuhmDLQ8kaEoyMjIFZJA=
modernc.org/mathutil v1.7.1 h1:GCZVGXdaN8gTqB1Mf/usp1Y/hSqgI2vAGGP4jZMCxOU=
modernc.org/mathutil v1.7.1/go.mod h1:4p5IwJITfppl0G4sUEDtCr4DthTaT47/N3aT6MhfgJg=
modernc.org/memory v1.11.0 h1:o4QC8aMQzmcwCK3t3Ux/ZHmwFPzE6hf2Y5LbkRs+hbI=
@@ -195,8 +197,8 @@ modernc.org/opt v0.1.4 h1:2kNGMRiUjrp4LcaPuLY2PzUfqM/w9N23quVwhKt5Qm8=
modernc.org/opt v0.1.4/go.mod h1:03fq9lsNfvkYSfxrfUhZCWPk1lm4cq4N+Bh//bEtgns=
modernc.org/sortutil v1.2.1 h1:+xyoGf15mM3NMlPDnFqrteY07klSFxLElE2PVuWIJ7w=
modernc.org/sortutil v1.2.1/go.mod h1:7ZI3a3REbai7gzCLcotuw9AC4VZVpYMjDzETGsSMqJE=
modernc.org/sqlite v1.45.0 h1:r51cSGzKpbptxnby+EIIz5fop4VuE4qFoVEjNvWoObs=
modernc.org/sqlite v1.45.0/go.mod h1:CzbrU2lSB1DKUusvwGz7rqEKIq+NUd8GWuBBZDs9/nA=
modernc.org/sqlite v1.40.0 h1:bNWEDlYhNPAUdUdBzjAvn8icAs/2gaKlj4vM+tQ6KdQ=
modernc.org/sqlite v1.40.0/go.mod h1:9fjQZ0mB1LLP0GYrp39oOJXx/I2sxEnZtzCmEQIKvGE=
modernc.org/strutil v1.2.1 h1:UneZBkQA+DX2Rp35KcM69cSsNES9ly8mQWD71HKlOA0=
modernc.org/strutil v1.2.1/go.mod h1:EHkiggD70koQxjVdSBM3JKM7k6L0FbGE5eymy9i3B9A=
modernc.org/token v1.1.0 h1:Xl7Ap9dKaEs5kLoOQeQmPWevfnk/DM5qcLcYlA8ys6Y=

View File

@@ -49,7 +49,6 @@ type SystemAlertStats struct {
GPU map[string]SystemAlertGPUData `json:"g"`
Temperatures map[string]float32 `json:"t"`
LoadAvg [3]float64 `json:"la"`
Battery [2]uint8 `json:"bat"`
}
type SystemAlertGPUData struct {
@@ -105,7 +104,6 @@ func NewAlertManager(app hubLike) *AlertManager {
func (am *AlertManager) bindEvents() {
am.hub.OnRecordAfterUpdateSuccess("alerts").BindFunc(updateHistoryOnAlertUpdate)
am.hub.OnRecordAfterDeleteSuccess("alerts").BindFunc(resolveHistoryOnAlertDelete)
am.hub.OnRecordAfterUpdateSuccess("smart_devices").BindFunc(am.handleSmartDeviceAlert)
}
// IsNotificationSilenced checks if a notification should be silenced based on configured quiet hours

View File

@@ -1,386 +0,0 @@
//go:build testing
package alerts_test
import (
"encoding/json"
"testing"
"time"
"github.com/henrygd/beszel/internal/entities/system"
beszelTests "github.com/henrygd/beszel/internal/tests"
"github.com/pocketbase/dbx"
"github.com/pocketbase/pocketbase/tools/types"
"github.com/stretchr/testify/assert"
"github.com/stretchr/testify/require"
)
// TestBatteryAlertLogic tests that battery alerts trigger when value drops BELOW threshold
// (opposite of other alerts like CPU, Memory, etc. which trigger when exceeding threshold)
func TestBatteryAlertLogic(t *testing.T) {
hub, user := beszelTests.GetHubWithUser(t)
defer hub.Cleanup()
// Create a system
systems, err := beszelTests.CreateSystems(hub, 1, user.Id, "up")
require.NoError(t, err)
systemRecord := systems[0]
// Create a battery alert with threshold of 20% and min of 1 minute (immediate trigger)
batteryAlert, err := beszelTests.CreateRecord(hub, "alerts", map[string]any{
"name": "Battery",
"system": systemRecord.Id,
"user": user.Id,
"value": 20, // threshold: 20%
"min": 1, // 1 minute (immediate trigger for testing)
})
require.NoError(t, err)
// Verify alert is not triggered initially
assert.False(t, batteryAlert.GetBool("triggered"), "Alert should not be triggered initially")
// Create system stats with battery at 50% (above threshold - should NOT trigger)
statsHigh := system.Stats{
Cpu: 10,
MemPct: 30,
DiskPct: 40,
Battery: [2]uint8{50, 1}, // 50% battery, discharging
}
statsHighJSON, _ := json.Marshal(statsHigh)
_, err = beszelTests.CreateRecord(hub, "system_stats", map[string]any{
"system": systemRecord.Id,
"type": "1m",
"stats": string(statsHighJSON),
})
require.NoError(t, err)
// Create CombinedData for the alert handler
combinedDataHigh := &system.CombinedData{
Stats: statsHigh,
Info: system.Info{
AgentVersion: "0.12.0",
Cpu: 10,
MemPct: 30,
DiskPct: 40,
},
}
// Simulate system update time
systemRecord.Set("updated", time.Now().UTC())
err = hub.SaveNoValidate(systemRecord)
require.NoError(t, err)
// Handle system alerts with high battery
am := hub.GetAlertManager()
err = am.HandleSystemAlerts(systemRecord, combinedDataHigh)
require.NoError(t, err)
// Verify alert is still NOT triggered (battery 50% is above threshold 20%)
batteryAlert, err = hub.FindFirstRecordByFilter("alerts", "id={:id}", dbx.Params{"id": batteryAlert.Id})
require.NoError(t, err)
assert.False(t, batteryAlert.GetBool("triggered"), "Alert should NOT be triggered when battery (50%%) is above threshold (20%%)")
// Now create stats with battery at 15% (below threshold - should trigger)
statsLow := system.Stats{
Cpu: 10,
MemPct: 30,
DiskPct: 40,
Battery: [2]uint8{15, 1}, // 15% battery, discharging
}
statsLowJSON, _ := json.Marshal(statsLow)
_, err = beszelTests.CreateRecord(hub, "system_stats", map[string]any{
"system": systemRecord.Id,
"type": "1m",
"stats": string(statsLowJSON),
})
require.NoError(t, err)
combinedDataLow := &system.CombinedData{
Stats: statsLow,
Info: system.Info{
AgentVersion: "0.12.0",
Cpu: 10,
MemPct: 30,
DiskPct: 40,
},
}
// Update system timestamp
systemRecord.Set("updated", time.Now().UTC())
err = hub.SaveNoValidate(systemRecord)
require.NoError(t, err)
// Handle system alerts with low battery
err = am.HandleSystemAlerts(systemRecord, combinedDataLow)
require.NoError(t, err)
// Wait for the alert to be processed
time.Sleep(20 * time.Millisecond)
// Verify alert IS triggered (battery 15% is below threshold 20%)
batteryAlert, err = hub.FindFirstRecordByFilter("alerts", "id={:id}", dbx.Params{"id": batteryAlert.Id})
require.NoError(t, err)
assert.True(t, batteryAlert.GetBool("triggered"), "Alert SHOULD be triggered when battery (15%%) drops below threshold (20%%)")
// Now test resolution: battery goes back above threshold
statsRecovered := system.Stats{
Cpu: 10,
MemPct: 30,
DiskPct: 40,
Battery: [2]uint8{25, 1}, // 25% battery, discharging
}
statsRecoveredJSON, _ := json.Marshal(statsRecovered)
_, err = beszelTests.CreateRecord(hub, "system_stats", map[string]any{
"system": systemRecord.Id,
"type": "1m",
"stats": string(statsRecoveredJSON),
})
require.NoError(t, err)
combinedDataRecovered := &system.CombinedData{
Stats: statsRecovered,
Info: system.Info{
AgentVersion: "0.12.0",
Cpu: 10,
MemPct: 30,
DiskPct: 40,
},
}
// Update system timestamp
systemRecord.Set("updated", time.Now().UTC())
err = hub.SaveNoValidate(systemRecord)
require.NoError(t, err)
// Handle system alerts with recovered battery
err = am.HandleSystemAlerts(systemRecord, combinedDataRecovered)
require.NoError(t, err)
// Wait for the alert to be processed
time.Sleep(20 * time.Millisecond)
// Verify alert is now resolved (battery 25% is above threshold 20%)
batteryAlert, err = hub.FindFirstRecordByFilter("alerts", "id={:id}", dbx.Params{"id": batteryAlert.Id})
require.NoError(t, err)
assert.False(t, batteryAlert.GetBool("triggered"), "Alert should be resolved when battery (25%%) goes above threshold (20%%)")
}
// TestBatteryAlertNoBattery verifies that systems without battery data don't trigger alerts
func TestBatteryAlertNoBattery(t *testing.T) {
hub, user := beszelTests.GetHubWithUser(t)
defer hub.Cleanup()
// Create a system
systems, err := beszelTests.CreateSystems(hub, 1, user.Id, "up")
require.NoError(t, err)
systemRecord := systems[0]
// Create a battery alert
batteryAlert, err := beszelTests.CreateRecord(hub, "alerts", map[string]any{
"name": "Battery",
"system": systemRecord.Id,
"user": user.Id,
"value": 20,
"min": 1,
})
require.NoError(t, err)
// Create stats with NO battery data (Battery[0] = 0)
statsNoBattery := system.Stats{
Cpu: 10,
MemPct: 30,
DiskPct: 40,
Battery: [2]uint8{0, 0}, // No battery
}
combinedData := &system.CombinedData{
Stats: statsNoBattery,
Info: system.Info{
AgentVersion: "0.12.0",
Cpu: 10,
MemPct: 30,
DiskPct: 40,
},
}
// Simulate system update time
systemRecord.Set("updated", time.Now().UTC())
err = hub.SaveNoValidate(systemRecord)
require.NoError(t, err)
// Handle system alerts
am := hub.GetAlertManager()
err = am.HandleSystemAlerts(systemRecord, combinedData)
require.NoError(t, err)
// Wait a moment for processing
time.Sleep(20 * time.Millisecond)
// Verify alert is NOT triggered (no battery data should skip the alert)
batteryAlert, err = hub.FindFirstRecordByFilter("alerts", "id={:id}", dbx.Params{"id": batteryAlert.Id})
require.NoError(t, err)
assert.False(t, batteryAlert.GetBool("triggered"), "Alert should NOT be triggered when system has no battery")
}
// TestBatteryAlertAveragedSamples tests battery alerts with min > 1 (averaging multiple samples)
// This ensures the inverted threshold logic works correctly across averaged time windows
func TestBatteryAlertAveragedSamples(t *testing.T) {
hub, user := beszelTests.GetHubWithUser(t)
defer hub.Cleanup()
// Create a system
systems, err := beszelTests.CreateSystems(hub, 1, user.Id, "up")
require.NoError(t, err)
systemRecord := systems[0]
// Create a battery alert with threshold of 25% and min of 2 minutes (requires averaging)
batteryAlert, err := beszelTests.CreateRecord(hub, "alerts", map[string]any{
"name": "Battery",
"system": systemRecord.Id,
"user": user.Id,
"value": 25, // threshold: 25%
"min": 2, // 2 minutes - requires averaging
})
require.NoError(t, err)
// Verify alert is not triggered initially
assert.False(t, batteryAlert.GetBool("triggered"), "Alert should not be triggered initially")
am := hub.GetAlertManager()
now := time.Now().UTC()
// Create system_stats records with low battery (below threshold)
// The alert has min=2 minutes, so alert.time = now - 2 minutes
// For the alert to be valid, alert.time must be AFTER the oldest record's created time
// So we need records older than (now - 2 min), plus records within the window
// Records at: now-3min (oldest, before window), now-90s, now-60s, now-30s
recordTimes := []time.Duration{
-180 * time.Second, // 3 min ago - this makes the oldest record before alert.time
-90 * time.Second,
-60 * time.Second,
-30 * time.Second,
}
for _, offset := range recordTimes {
statsLow := system.Stats{
Cpu: 10,
MemPct: 30,
DiskPct: 40,
Battery: [2]uint8{15, 1}, // 15% battery (below 25% threshold)
}
statsLowJSON, _ := json.Marshal(statsLow)
recordTime := now.Add(offset)
record, err := beszelTests.CreateRecord(hub, "system_stats", map[string]any{
"system": systemRecord.Id,
"type": "1m",
"stats": string(statsLowJSON),
})
require.NoError(t, err)
// Update created time to simulate historical records - use SetRaw with formatted string
record.SetRaw("created", recordTime.Format(types.DefaultDateLayout))
err = hub.SaveNoValidate(record)
require.NoError(t, err)
}
// Create combined data with low battery
combinedDataLow := &system.CombinedData{
Stats: system.Stats{
Cpu: 10,
MemPct: 30,
DiskPct: 40,
Battery: [2]uint8{15, 1},
},
Info: system.Info{
AgentVersion: "0.12.0",
Cpu: 10,
MemPct: 30,
DiskPct: 40,
},
}
// Update system timestamp
systemRecord.Set("updated", now)
err = hub.SaveNoValidate(systemRecord)
require.NoError(t, err)
// Handle system alerts - should trigger because average battery is below threshold
err = am.HandleSystemAlerts(systemRecord, combinedDataLow)
require.NoError(t, err)
// Wait for alert processing
time.Sleep(20 * time.Millisecond)
// Verify alert IS triggered (average battery 15% is below threshold 25%)
batteryAlert, err = hub.FindFirstRecordByFilter("alerts", "id={:id}", dbx.Params{"id": batteryAlert.Id})
require.NoError(t, err)
assert.True(t, batteryAlert.GetBool("triggered"),
"Alert SHOULD be triggered when average battery (15%%) is below threshold (25%%) over min period")
// Now add records with high battery to test resolution
// Use a new time window 2 minutes later
newNow := now.Add(2 * time.Minute)
// Records need to span before the alert time window (newNow - 2 min)
recordTimesHigh := []time.Duration{
-180 * time.Second, // 3 min before newNow - makes oldest record before alert.time
-90 * time.Second,
-60 * time.Second,
-30 * time.Second,
}
for _, offset := range recordTimesHigh {
statsHigh := system.Stats{
Cpu: 10,
MemPct: 30,
DiskPct: 40,
Battery: [2]uint8{50, 1}, // 50% battery (above 25% threshold)
}
statsHighJSON, _ := json.Marshal(statsHigh)
recordTime := newNow.Add(offset)
record, err := beszelTests.CreateRecord(hub, "system_stats", map[string]any{
"system": systemRecord.Id,
"type": "1m",
"stats": string(statsHighJSON),
})
require.NoError(t, err)
record.SetRaw("created", recordTime.Format(types.DefaultDateLayout))
err = hub.SaveNoValidate(record)
require.NoError(t, err)
}
// Create combined data with high battery
combinedDataHigh := &system.CombinedData{
Stats: system.Stats{
Cpu: 10,
MemPct: 30,
DiskPct: 40,
Battery: [2]uint8{50, 1},
},
Info: system.Info{
AgentVersion: "0.12.0",
Cpu: 10,
MemPct: 30,
DiskPct: 40,
},
}
// Update system timestamp to the new time window
systemRecord.Set("updated", newNow)
err = hub.SaveNoValidate(systemRecord)
require.NoError(t, err)
// Handle system alerts - should resolve because average battery is now above threshold
err = am.HandleSystemAlerts(systemRecord, combinedDataHigh)
require.NoError(t, err)
// Wait for alert processing
time.Sleep(20 * time.Millisecond)
// Verify alert is resolved (average battery 50% is above threshold 25%)
batteryAlert, err = hub.FindFirstRecordByFilter("alerts", "id={:id}", dbx.Params{"id": batteryAlert.Id})
require.NoError(t, err)
assert.False(t, batteryAlert.GetBool("triggered"),
"Alert should be resolved when average battery (50%%) is above threshold (25%%) over min period")
}

View File

@@ -1,4 +1,5 @@
//go:build testing
// +build testing
package alerts_test

View File

@@ -1,107 +0,0 @@
package alerts
import (
"fmt"
"strings"
"github.com/pocketbase/pocketbase/core"
)
// handleSmartDeviceAlert sends alerts when a SMART device state worsens into WARNING/FAILED.
// This is automatic and does not require user opt-in.
func (am *AlertManager) handleSmartDeviceAlert(e *core.RecordEvent) error {
oldState := e.Record.Original().GetString("state")
newState := e.Record.GetString("state")
if !shouldSendSmartDeviceAlert(oldState, newState) {
return e.Next()
}
systemID := e.Record.GetString("system")
if systemID == "" {
return e.Next()
}
// Fetch the system record to get the name and users
systemRecord, err := e.App.FindRecordById("systems", systemID)
if err != nil {
e.App.Logger().Error("Failed to find system for SMART alert", "err", err, "systemID", systemID)
return e.Next()
}
systemName := systemRecord.GetString("name")
deviceName := e.Record.GetString("name")
model := e.Record.GetString("model")
statusLabel := smartStateLabel(newState)
// Build alert message
title := fmt.Sprintf("SMART %s on %s: %s %s", statusLabel, systemName, deviceName, smartStateEmoji(newState))
var message string
if model != "" {
message = fmt.Sprintf("Disk %s (%s) SMART status changed to %s", deviceName, model, newState)
} else {
message = fmt.Sprintf("Disk %s SMART status changed to %s", deviceName, newState)
}
// Get users associated with the system
userIDs := systemRecord.GetStringSlice("users")
if len(userIDs) == 0 {
return e.Next()
}
// Send alert to each user
for _, userID := range userIDs {
if err := am.SendAlert(AlertMessageData{
UserID: userID,
SystemID: systemID,
Title: title,
Message: message,
Link: am.hub.MakeLink("system", systemID),
LinkText: "View " + systemName,
}); err != nil {
e.App.Logger().Error("Failed to send SMART alert", "err", err, "userID", userID)
}
}
return e.Next()
}
func shouldSendSmartDeviceAlert(oldState, newState string) bool {
oldSeverity := smartStateSeverity(oldState)
newSeverity := smartStateSeverity(newState)
// Ignore unknown states and recoveries; only alert on worsening transitions
// from known-good/degraded states into WARNING/FAILED.
return oldSeverity >= 1 && newSeverity > oldSeverity
}
func smartStateSeverity(state string) int {
switch state {
case "PASSED":
return 1
case "WARNING":
return 2
case "FAILED":
return 3
default:
return 0
}
}
func smartStateEmoji(state string) string {
switch state {
case "WARNING":
return "\U0001F7E0"
default:
return "\U0001F534"
}
}
func smartStateLabel(state string) string {
switch state {
case "FAILED":
return "failure"
default:
return strings.ToLower(state)
}
}

View File

@@ -1,264 +0,0 @@
//go:build testing
package alerts_test
import (
"testing"
"time"
beszelTests "github.com/henrygd/beszel/internal/tests"
"github.com/stretchr/testify/assert"
)
func TestSmartDeviceAlert(t *testing.T) {
hub, user := beszelTests.GetHubWithUser(t)
defer hub.Cleanup()
// Create a system for the user
system, err := beszelTests.CreateRecord(hub, "systems", map[string]any{
"name": "test-system",
"users": []string{user.Id},
"host": "127.0.0.1",
})
assert.NoError(t, err)
// Create a smart_device with state PASSED
smartDevice, err := beszelTests.CreateRecord(hub, "smart_devices", map[string]any{
"system": system.Id,
"name": "/dev/sda",
"model": "Samsung SSD 970 EVO",
"state": "PASSED",
})
assert.NoError(t, err)
// Verify no emails sent initially
assert.Zero(t, hub.TestMailer.TotalSend(), "should have 0 emails sent initially")
// Re-fetch the record so PocketBase can properly track original values
smartDevice, err = hub.FindRecordById("smart_devices", smartDevice.Id)
assert.NoError(t, err)
// Update the smart device state to FAILED
smartDevice.Set("state", "FAILED")
err = hub.Save(smartDevice)
assert.NoError(t, err)
// Wait for the alert to be processed
time.Sleep(50 * time.Millisecond)
// Verify that an email was sent
assert.EqualValues(t, 1, hub.TestMailer.TotalSend(), "should have 1 email sent after state changed to FAILED")
// Check the email content
lastMessage := hub.TestMailer.LastMessage()
assert.Contains(t, lastMessage.Subject, "SMART failure on test-system")
assert.Contains(t, lastMessage.Subject, "/dev/sda")
assert.Contains(t, lastMessage.Text, "Samsung SSD 970 EVO")
assert.Contains(t, lastMessage.Text, "FAILED")
}
func TestSmartDeviceAlertPassedToWarning(t *testing.T) {
hub, user := beszelTests.GetHubWithUser(t)
defer hub.Cleanup()
system, err := beszelTests.CreateRecord(hub, "systems", map[string]any{
"name": "test-system",
"users": []string{user.Id},
"host": "127.0.0.1",
})
assert.NoError(t, err)
smartDevice, err := beszelTests.CreateRecord(hub, "smart_devices", map[string]any{
"system": system.Id,
"name": "/dev/mmcblk0",
"model": "eMMC",
"state": "PASSED",
})
assert.NoError(t, err)
smartDevice, err = hub.FindRecordById("smart_devices", smartDevice.Id)
assert.NoError(t, err)
smartDevice.Set("state", "WARNING")
err = hub.Save(smartDevice)
assert.NoError(t, err)
time.Sleep(50 * time.Millisecond)
assert.EqualValues(t, 1, hub.TestMailer.TotalSend(), "should have 1 email sent after state changed to WARNING")
lastMessage := hub.TestMailer.LastMessage()
assert.Contains(t, lastMessage.Subject, "SMART warning on test-system")
assert.Contains(t, lastMessage.Text, "WARNING")
}
func TestSmartDeviceAlertWarningToFailed(t *testing.T) {
hub, user := beszelTests.GetHubWithUser(t)
defer hub.Cleanup()
system, err := beszelTests.CreateRecord(hub, "systems", map[string]any{
"name": "test-system",
"users": []string{user.Id},
"host": "127.0.0.1",
})
assert.NoError(t, err)
smartDevice, err := beszelTests.CreateRecord(hub, "smart_devices", map[string]any{
"system": system.Id,
"name": "/dev/mmcblk0",
"model": "eMMC",
"state": "WARNING",
})
assert.NoError(t, err)
smartDevice, err = hub.FindRecordById("smart_devices", smartDevice.Id)
assert.NoError(t, err)
smartDevice.Set("state", "FAILED")
err = hub.Save(smartDevice)
assert.NoError(t, err)
time.Sleep(50 * time.Millisecond)
assert.EqualValues(t, 1, hub.TestMailer.TotalSend(), "should have 1 email sent after state changed from WARNING to FAILED")
lastMessage := hub.TestMailer.LastMessage()
assert.Contains(t, lastMessage.Subject, "SMART failure on test-system")
assert.Contains(t, lastMessage.Text, "FAILED")
}
func TestSmartDeviceAlertNoAlertOnNonPassedToFailed(t *testing.T) {
hub, user := beszelTests.GetHubWithUser(t)
defer hub.Cleanup()
// Create a system for the user
system, err := beszelTests.CreateRecord(hub, "systems", map[string]any{
"name": "test-system",
"users": []string{user.Id},
"host": "127.0.0.1",
})
assert.NoError(t, err)
// Create a smart_device with state UNKNOWN
smartDevice, err := beszelTests.CreateRecord(hub, "smart_devices", map[string]any{
"system": system.Id,
"name": "/dev/sda",
"model": "Samsung SSD 970 EVO",
"state": "UNKNOWN",
})
assert.NoError(t, err)
// Re-fetch the record so PocketBase can properly track original values
smartDevice, err = hub.FindRecordById("smart_devices", smartDevice.Id)
assert.NoError(t, err)
// Update the state from UNKNOWN to FAILED - should NOT trigger alert.
// We only alert from known healthy/degraded states.
smartDevice.Set("state", "FAILED")
err = hub.Save(smartDevice)
assert.NoError(t, err)
time.Sleep(50 * time.Millisecond)
// Verify no email was sent (only PASSED -> FAILED triggers alert)
assert.Zero(t, hub.TestMailer.TotalSend(), "should have 0 emails when changing from UNKNOWN to FAILED")
// Re-fetch the record again
smartDevice, err = hub.FindRecordById("smart_devices", smartDevice.Id)
assert.NoError(t, err)
// Update state from FAILED to PASSED - should NOT trigger alert
smartDevice.Set("state", "PASSED")
err = hub.Save(smartDevice)
assert.NoError(t, err)
time.Sleep(50 * time.Millisecond)
// Verify no email was sent
assert.Zero(t, hub.TestMailer.TotalSend(), "should have 0 emails when changing from FAILED to PASSED")
}
func TestSmartDeviceAlertMultipleUsers(t *testing.T) {
hub, user1 := beszelTests.GetHubWithUser(t)
defer hub.Cleanup()
// Create a second user
user2, err := beszelTests.CreateUser(hub, "test2@example.com", "password")
assert.NoError(t, err)
// Create user settings for the second user
_, err = beszelTests.CreateRecord(hub, "user_settings", map[string]any{
"user": user2.Id,
"settings": `{"emails":["test2@example.com"],"webhooks":[]}`,
})
assert.NoError(t, err)
// Create a system with both users
system, err := beszelTests.CreateRecord(hub, "systems", map[string]any{
"name": "shared-system",
"users": []string{user1.Id, user2.Id},
"host": "127.0.0.1",
})
assert.NoError(t, err)
// Create a smart_device with state PASSED
smartDevice, err := beszelTests.CreateRecord(hub, "smart_devices", map[string]any{
"system": system.Id,
"name": "/dev/nvme0n1",
"model": "WD Black SN850",
"state": "PASSED",
})
assert.NoError(t, err)
// Re-fetch the record so PocketBase can properly track original values
smartDevice, err = hub.FindRecordById("smart_devices", smartDevice.Id)
assert.NoError(t, err)
// Update the smart device state to FAILED
smartDevice.Set("state", "FAILED")
err = hub.Save(smartDevice)
assert.NoError(t, err)
time.Sleep(50 * time.Millisecond)
// Verify that two emails were sent (one for each user)
assert.EqualValues(t, 2, hub.TestMailer.TotalSend(), "should have 2 emails sent for 2 users")
}
func TestSmartDeviceAlertWithoutModel(t *testing.T) {
hub, user := beszelTests.GetHubWithUser(t)
defer hub.Cleanup()
// Create a system for the user
system, err := beszelTests.CreateRecord(hub, "systems", map[string]any{
"name": "test-system",
"users": []string{user.Id},
"host": "127.0.0.1",
})
assert.NoError(t, err)
// Create a smart_device with state PASSED but no model
smartDevice, err := beszelTests.CreateRecord(hub, "smart_devices", map[string]any{
"system": system.Id,
"name": "/dev/sdb",
"state": "PASSED",
})
assert.NoError(t, err)
// Re-fetch the record so PocketBase can properly track original values
smartDevice, err = hub.FindRecordById("smart_devices", smartDevice.Id)
assert.NoError(t, err)
// Update the smart device state to FAILED
smartDevice.Set("state", "FAILED")
err = hub.Save(smartDevice)
assert.NoError(t, err)
time.Sleep(50 * time.Millisecond)
// Verify that an email was sent
assert.EqualValues(t, 1, hub.TestMailer.TotalSend(), "should have 1 email sent")
// Check that the email doesn't have empty parentheses for missing model
lastMessage := hub.TestMailer.LastMessage()
assert.NotContains(t, lastMessage.Text, "()", "should not have empty parentheses for missing model")
assert.Contains(t, lastMessage.Text, "/dev/sdb")
}

View File

@@ -66,30 +66,17 @@ func (am *AlertManager) HandleSystemAlerts(systemRecord *core.Record, data *syst
unit = ""
case "GPU":
val = data.Info.GpuPct
case "Battery":
if data.Stats.Battery[0] == 0 {
continue
}
val = float64(data.Stats.Battery[0])
}
triggered := alertRecord.GetBool("triggered")
threshold := alertRecord.GetFloat("value")
// Battery alert has inverted logic: trigger when value is BELOW threshold
lowAlert := isLowAlert(name)
// CONTINUE
// For normal alerts: IF not triggered and curValue <= threshold, OR triggered and curValue > threshold
// For low alerts (Battery): IF not triggered and curValue >= threshold, OR triggered and curValue < threshold
if lowAlert {
if (!triggered && val >= threshold) || (triggered && val < threshold) {
continue
}
} else {
if (!triggered && val <= threshold) || (triggered && val > threshold) {
continue
}
// IF alert is not triggered and curValue is less than threshold
// OR alert is triggered and curValue is greater than threshold
if (!triggered && val <= threshold) || (triggered && val > threshold) {
// log.Printf("Skipping alert %s: val %f | threshold %f | triggered %v\n", name, val, threshold, triggered)
continue
}
min := max(1, cast.ToUint8(alertRecord.Get("min")))
@@ -107,11 +94,7 @@ func (am *AlertManager) HandleSystemAlerts(systemRecord *core.Record, data *syst
// send alert immediately if min is 1 - no need to sum up values.
if min == 1 {
if lowAlert {
alert.triggered = val < threshold
} else {
alert.triggered = val > threshold
}
alert.triggered = val > threshold
go am.sendSystemAlert(alert)
continue
}
@@ -236,8 +219,6 @@ func (am *AlertManager) HandleSystemAlerts(systemRecord *core.Record, data *syst
}
}
alert.val += maxUsage
case "Battery":
alert.val += float64(stats.Battery[0])
default:
continue
}
@@ -275,24 +256,12 @@ func (am *AlertManager) HandleSystemAlerts(systemRecord *core.Record, data *syst
// log.Printf("%s: val %f | count %d | min-count %f | threshold %f\n", alert.name, alert.val, alert.count, minCount, alert.threshold)
// pass through alert if count is greater than or equal to minCount
if float32(alert.count) >= minCount {
// Battery alert has inverted logic: trigger when value is BELOW threshold
lowAlert := isLowAlert(alert.name)
if lowAlert {
if !alert.triggered && alert.val < alert.threshold {
alert.triggered = true
go am.sendSystemAlert(alert)
} else if alert.triggered && alert.val >= alert.threshold {
alert.triggered = false
go am.sendSystemAlert(alert)
}
} else {
if !alert.triggered && alert.val > alert.threshold {
alert.triggered = true
go am.sendSystemAlert(alert)
} else if alert.triggered && alert.val <= alert.threshold {
alert.triggered = false
go am.sendSystemAlert(alert)
}
if !alert.triggered && alert.val > alert.threshold {
alert.triggered = true
go am.sendSystemAlert(alert)
} else if alert.triggered && alert.val <= alert.threshold {
alert.triggered = false
go am.sendSystemAlert(alert)
}
}
}
@@ -319,19 +288,10 @@ func (am *AlertManager) sendSystemAlert(alert SystemAlertData) {
}
var subject string
lowAlert := isLowAlert(alert.name)
if alert.triggered {
if lowAlert {
subject = fmt.Sprintf("%s %s below threshold", systemName, titleAlertName)
} else {
subject = fmt.Sprintf("%s %s above threshold", systemName, titleAlertName)
}
subject = fmt.Sprintf("%s %s above threshold", systemName, titleAlertName)
} else {
if lowAlert {
subject = fmt.Sprintf("%s %s above threshold", systemName, titleAlertName)
} else {
subject = fmt.Sprintf("%s %s below threshold", systemName, titleAlertName)
}
subject = fmt.Sprintf("%s %s below threshold", systemName, titleAlertName)
}
minutesLabel := "minute"
if alert.min > 1 {
@@ -356,7 +316,3 @@ func (am *AlertManager) sendSystemAlert(alert SystemAlertData) {
LinkText: "View " + systemName,
})
}
func isLowAlert(name string) bool {
return name == "Battery"
}

View File

@@ -1,4 +1,5 @@
//go:build testing
// +build testing
package alerts_test

View File

@@ -1,4 +1,5 @@
//go:build testing
// +build testing
package alerts

View File

@@ -17,8 +17,9 @@ import (
type cmdOptions struct {
key string // key is the public key(s) for SSH authentication.
listen string // listen is the address or port to listen on.
hubURL string // hubURL is the URL of the Beszel hub.
token string // token is the token to use for authentication.
// TODO: add hubURL and token
// hubURL string // hubURL is the URL of the hub to use.
// token string // token is the token to use for authentication.
}
// parse parses the command line flags and populates the config struct.
@@ -31,6 +32,9 @@ func (opts *cmdOptions) parse() bool {
// Subcommands that don't require any pflag parsing
switch subcommand {
case "-v", "version":
fmt.Println(beszel.AppName+"-agent", beszel.Version)
return true
case "health":
err := health.Check()
if err != nil {
@@ -38,22 +42,18 @@ func (opts *cmdOptions) parse() bool {
}
fmt.Print("ok")
return true
case "fingerprint":
handleFingerprint()
return true
}
// pflag.CommandLine.ParseErrorsWhitelist.UnknownFlags = true
pflag.StringVarP(&opts.key, "key", "k", "", "Public key(s) for SSH authentication")
pflag.StringVarP(&opts.listen, "listen", "l", "", "Address or port to listen on")
pflag.StringVarP(&opts.hubURL, "url", "u", "", "URL of the Beszel hub")
pflag.StringVarP(&opts.token, "token", "t", "", "Token to use for authentication")
// pflag.StringVarP(&opts.hubURL, "hub-url", "u", "", "URL of the hub to use")
// pflag.StringVarP(&opts.token, "token", "t", "", "Token to use for authentication")
chinaMirrors := pflag.BoolP("china-mirrors", "c", false, "Use mirror for update (gh.beszel.dev) instead of GitHub")
version := pflag.BoolP("version", "v", false, "Show version information")
help := pflag.BoolP("help", "h", false, "Show this help message")
// Convert old single-dash long flags to double-dash for backward compatibility
flagsToConvert := []string{"key", "listen", "url", "token"}
flagsToConvert := []string{"key", "listen"}
for i, arg := range os.Args {
for _, flag := range flagsToConvert {
singleDash := "-" + flag
@@ -74,9 +74,9 @@ func (opts *cmdOptions) parse() bool {
builder.WriteString(os.Args[0])
builder.WriteString(" [command] [flags]\n")
builder.WriteString("\nCommands:\n")
builder.WriteString(" fingerprint View or reset the agent fingerprint\n")
builder.WriteString(" health Check if the agent is running\n")
builder.WriteString(" update Update to the latest version\n")
builder.WriteString(" health Check if the agent is running\n")
// builder.WriteString(" help Display this help message\n")
builder.WriteString(" update Update to the latest version\n")
builder.WriteString("\nFlags:\n")
fmt.Print(builder.String())
pflag.PrintDefaults()
@@ -87,9 +87,6 @@ func (opts *cmdOptions) parse() bool {
// Must run after pflag.Parse()
switch {
case *version:
fmt.Println(beszel.AppName+"-agent", beszel.Version)
return true
case *help || subcommand == "help":
pflag.Usage()
return true
@@ -98,13 +95,6 @@ func (opts *cmdOptions) parse() bool {
return true
}
// Set environment variables from CLI flags (if provided)
if opts.hubURL != "" {
os.Setenv("HUB_URL", opts.hubURL)
}
if opts.token != "" {
os.Setenv("TOKEN", opts.token)
}
return false
}
@@ -137,38 +127,6 @@ func (opts *cmdOptions) getAddress() string {
return agent.GetAddress(opts.listen)
}
// handleFingerprint handles the "fingerprint" command with subcommands "view" and "reset".
func handleFingerprint() {
subCmd := ""
if len(os.Args) > 2 {
subCmd = os.Args[2]
}
switch subCmd {
case "", "view":
dataDir, _ := agent.GetDataDir()
fp := agent.GetFingerprint(dataDir, "", "")
fmt.Println(fp)
case "help", "-h", "--help":
fmt.Print(fingerprintUsage())
case "reset":
dataDir, err := agent.GetDataDir()
if err != nil {
log.Fatal(err)
}
if err := agent.DeleteFingerprint(dataDir); err != nil {
log.Fatal(err)
}
fmt.Println("Fingerprint reset. A new one will be generated on next start.")
default:
log.Fatalf("Unknown command: %q\n\n%s", subCmd, fingerprintUsage())
}
}
func fingerprintUsage() string {
return fmt.Sprintf("Usage: %s fingerprint [view|reset]\n\nCommands:\n view Print fingerprint (default)\n reset Reset saved fingerprint\n", os.Args[0])
}
func main() {
var opts cmdOptions
subcommandHandled := opts.parse()

View File

@@ -1,7 +1,6 @@
package common
import (
"github.com/fxamacker/cbor/v2"
"github.com/henrygd/beszel/internal/entities/smart"
"github.com/henrygd/beszel/internal/entities/system"
"github.com/henrygd/beszel/internal/entities/systemd"
@@ -35,14 +34,14 @@ type HubRequest[T any] struct {
// AgentResponse defines the structure for responses sent from agent to hub.
type AgentResponse struct {
Id *uint32 `cbor:"0,keyasint,omitempty"`
SystemData *system.CombinedData `cbor:"1,keyasint,omitempty,omitzero"` // Legacy (<= 0.17)
Fingerprint *FingerprintResponse `cbor:"2,keyasint,omitempty,omitzero"` // Legacy (<= 0.17)
SystemData *system.CombinedData `cbor:"1,keyasint,omitempty,omitzero"`
Fingerprint *FingerprintResponse `cbor:"2,keyasint,omitempty,omitzero"`
Error string `cbor:"3,keyasint,omitempty,omitzero"`
String *string `cbor:"4,keyasint,omitempty,omitzero"` // Legacy (<= 0.17)
SmartData map[string]smart.SmartData `cbor:"5,keyasint,omitempty,omitzero"` // Legacy (<= 0.17)
ServiceInfo systemd.ServiceDetails `cbor:"6,keyasint,omitempty,omitzero"` // Legacy (<= 0.17)
// Data is the generic response payload for new endpoints (0.18+)
Data cbor.RawMessage `cbor:"7,keyasint,omitempty,omitzero"`
String *string `cbor:"4,keyasint,omitempty,omitzero"`
SmartData map[string]smart.SmartData `cbor:"5,keyasint,omitempty,omitzero"`
ServiceInfo systemd.ServiceDetails `cbor:"6,keyasint,omitempty,omitzero"`
// Logs *LogsPayload `cbor:"4,keyasint,omitempty,omitzero"`
// RawBytes []byte `cbor:"4,keyasint,omitempty,omitzero"`
}
type FingerprintRequest struct {
@@ -59,8 +58,8 @@ type FingerprintResponse struct {
}
type DataRequestOptions struct {
CacheTimeMs uint16 `cbor:"0,keyasint"`
IncludeDetails bool `cbor:"1,keyasint"`
CacheTimeMs uint16 `cbor:"0,keyasint"`
// ResourceType uint8 `cbor:"1,keyasint,omitempty,omitzero"`
}
type ContainerLogsRequest struct {

View File

@@ -23,9 +23,6 @@ COPY --from=builder /agent /agent
# this is so we don't need to create the /tmp directory in the scratch container
COPY --from=builder /tmp /tmp
# AMD GPU name lookup (used by agent on Linux when /usr/share/libdrm/amdgpu.ids is read)
COPY --from=builder /app/agent/test-data/amdgpu.ids /usr/share/libdrm/amdgpu.ids
# Ensure data persistence across container recreations
VOLUME ["/var/lib/beszel-agent"]

View File

@@ -17,12 +17,9 @@ RUN rm -rf /tmp/*
# --------------------------
# Final image: default scratch-based agent
# --------------------------
FROM alpine:3.23
FROM alpine:latest
COPY --from=builder /agent /agent
# AMD GPU name lookup (used by agent on Linux when /usr/share/libdrm/amdgpu.ids is read)
COPY --from=builder /app/agent/test-data/amdgpu.ids /usr/share/libdrm/amdgpu.ids
RUN apk add --no-cache smartmontools
# Ensure data persistence across container recreations

View File

@@ -16,7 +16,7 @@ RUN CGO_ENABLED=0 GOGC=75 GOOS=$TARGETOS GOARCH=$TARGETARCH go build -ldflags "-
# Final image
# Note: must cap_add: [CAP_PERFMON] and mount /dev/dri/ as volume
# --------------------------
FROM alpine:3.23
FROM alpine:edge
COPY --from=builder /agent /agent

View File

@@ -1,4 +1,4 @@
FROM --platform=$BUILDPLATFORM golang:bookworm AS builder
FROM --platform=$BUILDPLATFORM golang:alpine AS builder
WORKDIR /app
@@ -10,7 +10,7 @@ COPY . ./
# Build
ARG TARGETOS TARGETARCH
RUN CGO_ENABLED=0 GOGC=75 GOOS=$TARGETOS GOARCH=$TARGETARCH go build -tags glibc -ldflags "-w -s" -o /agent ./internal/cmd/agent
RUN CGO_ENABLED=0 GOGC=75 GOOS=$TARGETOS GOARCH=$TARGETARCH go build -ldflags "-w -s" -o /agent ./internal/cmd/agent
# --------------------------
# Smartmontools builder stage
@@ -37,9 +37,6 @@ RUN apt-get update && apt-get install -y \
FROM nvidia/cuda:12.2.2-base-ubuntu22.04
COPY --from=builder /agent /agent
# AMD GPU name lookup (used by agent on hybrid laptops when /usr/share/libdrm/amdgpu.ids is read)
COPY --from=builder /app/agent/test-data/amdgpu.ids /usr/share/libdrm/amdgpu.ids
# Copy smartmontools binaries and config files
COPY --from=smartmontools-builder /usr/sbin/smartctl /usr/sbin/smartctl

View File

@@ -34,14 +34,6 @@ type ApiStats struct {
MemoryStats MemoryStats `json:"memory_stats"`
}
// Docker system info from /info API endpoint
type HostInfo struct {
OperatingSystem string `json:"OperatingSystem"`
KernelVersion string `json:"KernelVersion"`
NCPU int `json:"NCPU"`
MemTotal uint64 `json:"MemTotal"`
}
func (s *ApiStats) CalculateCpuPercentLinux(prevCpuContainer uint64, prevCpuSystem uint64) float64 {
cpuDelta := s.CPUStats.CPUUsage.TotalUsage - prevCpuContainer
systemDelta := s.CPUStats.SystemUsage - prevCpuSystem
@@ -129,12 +121,11 @@ var DockerHealthStrings = map[string]DockerHealth{
// Docker container stats
type Stats struct {
Name string `json:"n" cbor:"0,keyasint"`
Cpu float64 `json:"c" cbor:"1,keyasint"`
Mem float64 `json:"m" cbor:"2,keyasint"`
NetworkSent float64 `json:"ns,omitzero" cbor:"3,keyasint,omitzero"` // deprecated 0.18.3 (MB) - keep field for old agents/records
NetworkRecv float64 `json:"nr,omitzero" cbor:"4,keyasint,omitzero"` // deprecated 0.18.3 (MB) - keep field for old agents/records
Bandwidth [2]uint64 `json:"b,omitzero" cbor:"9,keyasint,omitzero"` // [sent bytes, recv bytes]
Name string `json:"n" cbor:"0,keyasint"`
Cpu float64 `json:"c" cbor:"1,keyasint"`
Mem float64 `json:"m" cbor:"2,keyasint"`
NetworkSent float64 `json:"ns" cbor:"3,keyasint"`
NetworkRecv float64 `json:"nr" cbor:"4,keyasint"`
Health DockerHealth `json:"-" cbor:"5,keyasint"`
Status string `json:"-" cbor:"6,keyasint"`

View File

@@ -130,23 +130,10 @@ type SummaryInfo struct {
}
type AtaSmartAttributes struct {
// Revision int `json:"revision"`
Table []AtaSmartAttribute `json:"table"`
}
type AtaDeviceStatistics struct {
Pages []AtaDeviceStatisticsPage `json:"pages"`
}
type AtaDeviceStatisticsPage struct {
Number uint8 `json:"number"`
Table []AtaDeviceStatisticsEntry `json:"table"`
}
type AtaDeviceStatisticsEntry struct {
Name string `json:"name"`
Value *uint64 `json:"value,omitempty"`
}
type AtaSmartAttribute struct {
ID uint16 `json:"id"`
Name string `json:"name"`
@@ -356,8 +343,7 @@ type SmartInfoForSata struct {
SmartStatus SmartStatusInfo `json:"smart_status"`
// AtaSmartData AtaSmartData `json:"ata_smart_data"`
// AtaSctCapabilities AtaSctCapabilities `json:"ata_sct_capabilities"`
AtaSmartAttributes AtaSmartAttributes `json:"ata_smart_attributes"`
AtaDeviceStatistics AtaDeviceStatistics `json:"ata_device_statistics"`
AtaSmartAttributes AtaSmartAttributes `json:"ata_smart_attributes"`
// PowerOnTime PowerOnTimeInfo `json:"power_on_time"`
// PowerCycleCount uint16 `json:"power_cycle_count"`
Temperature TemperatureInfo `json:"temperature"`

View File

@@ -27,8 +27,8 @@ type Stats struct {
DiskWritePs float64 `json:"dw" cbor:"13,keyasint"`
MaxDiskReadPs float64 `json:"drm,omitempty" cbor:"14,keyasint,omitempty"`
MaxDiskWritePs float64 `json:"dwm,omitempty" cbor:"15,keyasint,omitempty"`
NetworkSent float64 `json:"ns,omitzero" cbor:"16,keyasint,omitzero"`
NetworkRecv float64 `json:"nr,omitzero" cbor:"17,keyasint,omitzero"`
NetworkSent float64 `json:"ns" cbor:"16,keyasint"`
NetworkRecv float64 `json:"nr" cbor:"17,keyasint"`
MaxNetworkSent float64 `json:"nsm,omitempty" cbor:"18,keyasint,omitempty"`
MaxNetworkRecv float64 `json:"nrm,omitempty" cbor:"19,keyasint,omitempty"`
Temperatures map[string]float64 `json:"t,omitempty" cbor:"20,keyasint,omitempty"`
@@ -123,49 +123,31 @@ const (
ConnectionTypeWebSocket
)
// Core system data that is needed in All Systems table
type Info struct {
Hostname string `json:"h,omitempty" cbor:"0,keyasint,omitempty"` // deprecated - moved to Details struct
KernelVersion string `json:"k,omitempty" cbor:"1,keyasint,omitempty"` // deprecated - moved to Details struct
Cores int `json:"c,omitzero" cbor:"2,keyasint,omitzero"` // deprecated - moved to Details struct
// Threads is needed in Info struct to calculate load average thresholds
Hostname string `json:"h" cbor:"0,keyasint"`
KernelVersion string `json:"k,omitempty" cbor:"1,keyasint,omitempty"`
Cores int `json:"c" cbor:"2,keyasint"`
Threads int `json:"t,omitempty" cbor:"3,keyasint,omitempty"`
CpuModel string `json:"m,omitempty" cbor:"4,keyasint,omitempty"` // deprecated - moved to Details struct
CpuModel string `json:"m" cbor:"4,keyasint"`
Uptime uint64 `json:"u" cbor:"5,keyasint"`
Cpu float64 `json:"cpu" cbor:"6,keyasint"`
MemPct float64 `json:"mp" cbor:"7,keyasint"`
DiskPct float64 `json:"dp" cbor:"8,keyasint"`
Bandwidth float64 `json:"b" cbor:"9,keyasint"`
AgentVersion string `json:"v" cbor:"10,keyasint"`
Podman bool `json:"p,omitempty" cbor:"11,keyasint,omitempty"` // deprecated - moved to Details struct
Podman bool `json:"p,omitempty" cbor:"11,keyasint,omitempty"`
GpuPct float64 `json:"g,omitempty" cbor:"12,keyasint,omitempty"`
DashboardTemp float64 `json:"dt,omitempty" cbor:"13,keyasint,omitempty"`
Os Os `json:"os,omitempty" cbor:"14,keyasint,omitempty"` // deprecated - moved to Details struct
LoadAvg1 float64 `json:"l1,omitempty" cbor:"15,keyasint,omitempty"` // deprecated - use `la` array instead
LoadAvg5 float64 `json:"l5,omitempty" cbor:"16,keyasint,omitempty"` // deprecated - use `la` array instead
LoadAvg15 float64 `json:"l15,omitempty" cbor:"17,keyasint,omitempty"` // deprecated - use `la` array instead
Os Os `json:"os" cbor:"14,keyasint"`
LoadAvg1 float64 `json:"l1,omitempty" cbor:"15,keyasint,omitempty"`
LoadAvg5 float64 `json:"l5,omitempty" cbor:"16,keyasint,omitempty"`
LoadAvg15 float64 `json:"l15,omitempty" cbor:"17,keyasint,omitempty"`
BandwidthBytes uint64 `json:"bb" cbor:"18,keyasint"`
// TODO: remove load fields in future release in favor of load avg array
LoadAvg [3]float64 `json:"la,omitempty" cbor:"19,keyasint"`
ConnectionType ConnectionType `json:"ct,omitempty" cbor:"20,keyasint,omitempty,omitzero"`
ExtraFsPct map[string]float64 `json:"efs,omitempty" cbor:"21,keyasint,omitempty"`
Services []uint16 `json:"sv,omitempty" cbor:"22,keyasint,omitempty"` // [totalServices, numFailedServices]
Battery [2]uint8 `json:"bat,omitzero" cbor:"23,keyasint,omitzero"` // [percent, charge state]
}
// Data that does not change during process lifetime and is not needed in All Systems table
type Details struct {
Hostname string `cbor:"0,keyasint"`
Kernel string `cbor:"1,keyasint,omitempty"`
Cores int `cbor:"2,keyasint"`
Threads int `cbor:"3,keyasint"`
CpuModel string `cbor:"4,keyasint"`
Os Os `cbor:"5,keyasint"`
OsName string `cbor:"6,keyasint"`
Arch string `cbor:"7,keyasint"`
Podman bool `cbor:"8,keyasint,omitempty"`
MemoryTotal uint64 `cbor:"9,keyasint"`
SmartInterval time.Duration `cbor:"10,keyasint,omitempty"`
}
// Final data structure to return to the hub
@@ -174,5 +156,4 @@ type CombinedData struct {
Info Info `json:"info" cbor:"1,keyasint"`
Containers []*container.Stats `json:"container" cbor:"2,keyasint"`
SystemdServices []*systemd.Service `json:"systemd,omitempty" cbor:"3,keyasint,omitempty"`
Details *Details `cbor:"4,keyasint,omitempty"`
}

View File

@@ -11,7 +11,6 @@ import (
"log/slog"
"net/http"
"os"
"os/exec"
"path/filepath"
"runtime"
"strings"
@@ -34,7 +33,7 @@ func ColorPrint(color, text string) {
fmt.Println(color + text + colorReset)
}
func ColorPrintf(color, format string, args ...any) {
func ColorPrintf(color, format string, args ...interface{}) {
fmt.Printf(color+format+colorReset+"\n", args...)
}
@@ -346,32 +345,5 @@ func archiveSuffix(binaryName, goos, goarch string) string {
if goos == "windows" {
return fmt.Sprintf("%s_%s_%s.zip", binaryName, goos, goarch)
}
// Use glibc build for agent on glibc systems (includes NVML support via purego)
if binaryName == "beszel-agent" && goos == "linux" && goarch == "amd64" && isGlibc() {
return fmt.Sprintf("%s_%s_%s_glibc.tar.gz", binaryName, goos, goarch)
}
return fmt.Sprintf("%s_%s_%s.tar.gz", binaryName, goos, goarch)
}
func isGlibc() bool {
for _, path := range []string{
"/lib64/ld-linux-x86-64.so.2", // common on many distros
"/lib/x86_64-linux-gnu/ld-linux-x86-64.so.2", // Debian/Ubuntu
"/lib/ld-linux-x86-64.so.2", // alternate
} {
if _, err := os.Stat(path); err == nil {
return true
}
}
// Fallback to ldd output when present (musl ldd reports musl, glibc reports GNU libc/glibc).
if lddPath, err := exec.LookPath("ldd"); err == nil {
out, err := exec.Command(lddPath, "--version").CombinedOutput()
if err == nil {
s := strings.ToLower(string(out))
if strings.Contains(s, "gnu libc") || strings.Contains(s, "glibc") {
return true
}
}
}
return false
}

View File

@@ -1,66 +0,0 @@
package ghupdate
import (
"fmt"
"os/exec"
"strings"
)
// HandleSELinuxContext restores or applies the correct SELinux label to the binary.
func HandleSELinuxContext(path string) error {
out, err := exec.Command("getenforce").Output()
if err != nil {
// SELinux not enabled or getenforce not available
return nil
}
state := strings.TrimSpace(string(out))
if state == "Disabled" {
return nil
}
ColorPrint(ColorYellow, "SELinux is enabled; applying context…")
// Try persistent context via semanage+restorecon
if success := trySemanageRestorecon(path); success {
return nil
}
// Fallback to temporary context via chcon
if chconPath, err := exec.LookPath("chcon"); err == nil {
if err := exec.Command(chconPath, "-t", "bin_t", path).Run(); err != nil {
return fmt.Errorf("chcon failed: %w", err)
}
return nil
}
return fmt.Errorf("no SELinux tools available (semanage/restorecon or chcon)")
}
// trySemanageRestorecon attempts to set persistent SELinux context using semanage and restorecon.
// Returns true if successful, false otherwise.
func trySemanageRestorecon(path string) bool {
semanagePath, err := exec.LookPath("semanage")
if err != nil {
return false
}
restoreconPath, err := exec.LookPath("restorecon")
if err != nil {
return false
}
// Try to add the fcontext rule; if it already exists, try to modify it
if err := exec.Command(semanagePath, "fcontext", "-a", "-t", "bin_t", path).Run(); err != nil {
// Rule may already exist, try modify instead
if err := exec.Command(semanagePath, "fcontext", "-m", "-t", "bin_t", path).Run(); err != nil {
return false
}
}
// Apply the context with restorecon
if err := exec.Command(restoreconPath, "-v", path).Run(); err != nil {
return false
}
return true
}

View File

@@ -1,53 +0,0 @@
package ghupdate
import (
"os"
"os/exec"
"path/filepath"
"testing"
)
func TestHandleSELinuxContext_NoSELinux(t *testing.T) {
// Skip on SELinux systems - this test is for non-SELinux behavior
if _, err := exec.LookPath("getenforce"); err == nil {
t.Skip("skipping on SELinux-enabled system")
}
// On systems without SELinux, getenforce will fail and the function
// should return nil without error
tempFile := filepath.Join(t.TempDir(), "test-binary")
if err := os.WriteFile(tempFile, []byte("test"), 0755); err != nil {
t.Fatalf("failed to create temp file: %v", err)
}
err := HandleSELinuxContext(tempFile)
if err != nil {
t.Errorf("HandleSELinuxContext() on non-SELinux system returned error: %v", err)
}
}
func TestHandleSELinuxContext_InvalidPath(t *testing.T) {
// Skip on SELinux systems - this test is for non-SELinux behavior
if _, err := exec.LookPath("getenforce"); err == nil {
t.Skip("skipping on SELinux-enabled system")
}
// On non-SELinux systems, getenforce fails early so even invalid paths succeed
err := HandleSELinuxContext("/nonexistent/path/binary")
if err != nil {
t.Errorf("HandleSELinuxContext() with invalid path on non-SELinux system returned error: %v", err)
}
}
func TestTrySemanageRestorecon_NoTools(t *testing.T) {
// Skip if semanage is available (we don't want to modify system SELinux policy)
if _, err := exec.LookPath("semanage"); err == nil {
t.Skip("skipping on system with semanage available")
}
// Should return false when semanage is not available
result := trySemanageRestorecon("/some/path")
if result {
t.Error("trySemanageRestorecon() returned true when semanage is not available")
}
}

View File

@@ -66,15 +66,6 @@ func (acr *agentConnectRequest) agentConnect() (err error) {
// Check if token is an active universal token
acr.userId, acr.isUniversalToken = universalTokenMap.GetMap().GetOk(acr.token)
if !acr.isUniversalToken {
// Fallback: check for a permanent universal token stored in the DB
if rec, err := acr.hub.FindFirstRecordByFilter("universal_tokens", "token = {:token}", dbx.Params{"token": acr.token}); err == nil {
if userID := rec.GetString("user"); userID != "" {
acr.userId = userID
acr.isUniversalToken = true
}
}
}
// Find matching fingerprint records for this token
fpRecords := getFingerprintRecordsByToken(acr.token, acr.hub)

View File

@@ -1,4 +1,5 @@
//go:build testing
// +build testing
package hub
@@ -9,7 +10,6 @@ import (
"net/http/httptest"
"os"
"path/filepath"
"runtime"
"strings"
"testing"
"time"
@@ -35,26 +35,6 @@ func createTestHub(t testing.TB) (*Hub, *pbtests.TestApp, error) {
return NewHub(testApp), testApp, nil
}
// cleanupTestHub stops background system goroutines before tearing down the app.
func cleanupTestHub(hub *Hub, testApp *pbtests.TestApp) {
if hub != nil {
sm := hub.GetSystemManager()
sm.RemoveAllSystems()
// Give updater goroutines a brief window to observe cancellation before DB teardown.
for range 20 {
if sm.GetSystemCount() == 0 {
break
}
runtime.Gosched()
time.Sleep(5 * time.Millisecond)
}
time.Sleep(20 * time.Millisecond)
}
if testApp != nil {
testApp.Cleanup()
}
}
// Helper function to create a test record
func createTestRecord(app core.App, collection string, data map[string]any) (*core.Record, error) {
col, err := app.FindCachedCollectionByNameOrId(collection)
@@ -84,7 +64,7 @@ func TestValidateAgentHeaders(t *testing.T) {
if err != nil {
t.Fatal(err)
}
defer cleanupTestHub(hub, testApp)
defer testApp.Cleanup()
testCases := []struct {
name string
@@ -165,7 +145,7 @@ func TestGetAllFingerprintRecordsByToken(t *testing.T) {
if err != nil {
t.Fatal(err)
}
defer cleanupTestHub(hub, testApp)
defer testApp.Cleanup()
// create test user
userRecord, err := createTestUser(testApp)
@@ -255,7 +235,7 @@ func TestSetFingerprint(t *testing.T) {
if err != nil {
t.Fatal(err)
}
defer cleanupTestHub(hub, testApp)
defer testApp.Cleanup()
// Create test user
userRecord, err := createTestUser(testApp)
@@ -335,7 +315,7 @@ func TestCreateSystemFromAgentData(t *testing.T) {
if err != nil {
t.Fatal(err)
}
defer cleanupTestHub(hub, testApp)
defer testApp.Cleanup()
// Create test user
userRecord, err := createTestUser(testApp)
@@ -445,7 +425,7 @@ func TestUniversalTokenFlow(t *testing.T) {
if err != nil {
t.Fatal(err)
}
defer cleanupTestHub(nil, testApp)
defer testApp.Cleanup()
// Create test user
userRecord, err := createTestUser(testApp)
@@ -513,7 +493,7 @@ func TestAgentConnect(t *testing.T) {
if err != nil {
t.Fatal(err)
}
defer cleanupTestHub(hub, testApp)
defer testApp.Cleanup()
// Create test user
userRecord, err := createTestUser(testApp)
@@ -672,7 +652,7 @@ func TestHandleAgentConnect(t *testing.T) {
if err != nil {
t.Fatal(err)
}
defer cleanupTestHub(hub, testApp)
defer testApp.Cleanup()
// Create test user
userRecord, err := createTestUser(testApp)
@@ -757,7 +737,7 @@ func TestAgentWebSocketIntegration(t *testing.T) {
// Create hub and test app
hub, testApp, err := createTestHub(t)
require.NoError(t, err)
defer cleanupTestHub(hub, testApp)
defer testApp.Cleanup()
// Get the hub's SSH key
hubSigner, err := hub.GetSSHKey("")
@@ -962,8 +942,6 @@ func TestAgentWebSocketIntegration(t *testing.T) {
}
}
time.Sleep(20 * time.Millisecond)
// Verify fingerprint state by re-reading the specific record
updatedFingerprintRecord, err := testApp.FindRecordById("fingerprints", fingerprintRecord.Id)
require.NoError(t, err)
@@ -998,7 +976,7 @@ func TestMultipleSystemsWithSameUniversalToken(t *testing.T) {
// Create hub and test app
hub, testApp, err := createTestHub(t)
require.NoError(t, err)
defer cleanupTestHub(hub, testApp)
defer testApp.Cleanup()
// Get the hub's SSH key
hubSigner, err := hub.GetSSHKey("")
@@ -1166,8 +1144,6 @@ func TestMultipleSystemsWithSameUniversalToken(t *testing.T) {
assert.Equal(t, systemCount, systemsAfterCount, "Total system count should remain the same")
}
time.Sleep(20 * time.Millisecond)
// Verify that a fingerprint record exists for this fingerprint
fingerprints, err := testApp.FindRecordsByFilter("fingerprints", "token = {:token} && fingerprint = {:fingerprint}", "", -1, 0, map[string]any{
"token": universalToken,
@@ -1193,111 +1169,11 @@ func TestMultipleSystemsWithSameUniversalToken(t *testing.T) {
}
}
// TestPermanentUniversalTokenFromDB verifies that a universal token persisted in the DB
// (universal_tokens collection) is accepted for agent self-registration even if it is not
// present in the in-memory universalTokenMap.
func TestPermanentUniversalTokenFromDB(t *testing.T) {
// Create hub and test app
hub, testApp, err := createTestHub(t)
require.NoError(t, err)
defer cleanupTestHub(hub, testApp)
// Get the hub's SSH key
hubSigner, err := hub.GetSSHKey("")
require.NoError(t, err)
goodPubKey := hubSigner.PublicKey()
// Create test user
userRecord, err := createTestUser(testApp)
require.NoError(t, err)
// Create a permanent universal token record in the DB (do NOT add it to universalTokenMap)
universalToken := "db-universal-token-123"
_, err = createTestRecord(testApp, "universal_tokens", map[string]any{
"user": userRecord.Id,
"token": universalToken,
})
require.NoError(t, err)
// Create HTTP server with the actual API route
ts := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
if r.URL.Path == "/api/beszel/agent-connect" {
acr := &agentConnectRequest{
hub: hub,
req: r,
res: w,
}
acr.agentConnect()
} else {
http.NotFound(w, r)
}
}))
defer ts.Close()
// Create and configure agent
agentDataDir := t.TempDir()
err = os.WriteFile(filepath.Join(agentDataDir, "fingerprint"), []byte("db-token-system-fingerprint"), 0644)
require.NoError(t, err)
testAgent, err := agent.NewAgent(agentDataDir)
require.NoError(t, err)
// Set up environment variables for the agent
os.Setenv("BESZEL_AGENT_HUB_URL", ts.URL)
os.Setenv("BESZEL_AGENT_TOKEN", universalToken)
defer func() {
os.Unsetenv("BESZEL_AGENT_HUB_URL")
os.Unsetenv("BESZEL_AGENT_TOKEN")
}()
// Start agent in background
done := make(chan error, 1)
go func() {
serverOptions := agent.ServerOptions{
Network: "tcp",
Addr: "127.0.0.1:46050",
Keys: []ssh.PublicKey{goodPubKey},
}
done <- testAgent.Start(serverOptions)
}()
// Wait for connection result
maxWait := 2 * time.Second
time.Sleep(20 * time.Millisecond)
checkInterval := 20 * time.Millisecond
timeout := time.After(maxWait)
ticker := time.Tick(checkInterval)
connectionManager := testAgent.GetConnectionManager()
for {
select {
case <-timeout:
t.Fatalf("Expected connection to succeed but timed out - agent state: %d", connectionManager.State)
case <-ticker:
if connectionManager.State == agent.WebSocketConnected {
// Success
goto verify
}
case err := <-done:
// If Start returns early, treat it as failure
if err != nil {
t.Fatalf("Agent failed to start/connect: %v", err)
}
}
}
verify:
// Verify that a system was created for the user (self-registration path)
systemsAfter, err := testApp.FindRecordsByFilter("systems", "users ~ {:userId}", "", -1, 0, map[string]any{"userId": userRecord.Id})
require.NoError(t, err)
require.NotEmpty(t, systemsAfter, "Expected a system to be created for DB-backed universal token")
}
// TestFindOrCreateSystemForToken tests the findOrCreateSystemForToken function
func TestFindOrCreateSystemForToken(t *testing.T) {
hub, testApp, err := createTestHub(t)
require.NoError(t, err)
defer cleanupTestHub(hub, testApp)
defer testApp.Cleanup()
// Create test user
userRecord, err := createTestUser(testApp)

View File

@@ -1,4 +1,5 @@
//go:build testing
// +build testing
package config_test

View File

@@ -1,4 +1,5 @@
//go:build testing
// +build testing
package expirymap
@@ -414,11 +415,7 @@ func TestExpiryMap_RemoveValue_WithExpiration(t *testing.T) {
// Wait for first value to expire
time.Sleep(time.Millisecond * 20)
// Trigger lazy cleanup of the expired key
_, ok := em.GetOk("key1")
assert.False(t, ok)
// Try to remove the remaining "value1" entry (key3)
// Try to remove the expired value - should remove one of the "value1" entries
removedValue, ok := em.RemovebyValue("value1")
assert.True(t, ok)
assert.Equal(t, "value1", removedValue)
@@ -426,9 +423,14 @@ func TestExpiryMap_RemoveValue_WithExpiration(t *testing.T) {
// Should still have key2 (different value)
assert.True(t, em.Has("key2"))
// key1 should be gone due to expiration and key3 should be removed by value.
assert.False(t, em.Has("key1"))
assert.False(t, em.Has("key3"))
// Should have removed one of the "value1" entries (either key1 or key3)
// But we can't predict which one due to map iteration order
key1Exists := em.Has("key1")
key3Exists := em.Has("key3")
// Exactly one of key1 or key3 should be gone
assert.False(t, key1Exists && key3Exists) // Both shouldn't exist
assert.True(t, key1Exists || key3Exists) // At least one should still exist
}
func TestExpiryMap_ValueOperations_Integration(t *testing.T) {

View File

@@ -1,303 +0,0 @@
// Package heartbeat sends periodic outbound pings to an external monitoring
// endpoint (e.g. BetterStack, Uptime Kuma, Healthchecks.io) so operators can
// monitor Beszel without exposing it to the internet.
package heartbeat
import (
"bytes"
"encoding/json"
"fmt"
"net/http"
"net/url"
"strconv"
"strings"
"time"
"github.com/henrygd/beszel"
"github.com/pocketbase/pocketbase/core"
)
// Default values for heartbeat configuration.
const (
defaultInterval = 60 // seconds
httpTimeout = 10 * time.Second
)
// Payload is the JSON body sent with each heartbeat request.
type Payload struct {
// Status is "ok" when all non-paused systems are up, "warn" when alerts
// are triggered but no systems are down, and "error" when any system is down.
Status string `json:"status"`
Timestamp string `json:"timestamp"`
Msg string `json:"msg"`
Systems SystemsSummary `json:"systems"`
Down []SystemInfo `json:"down_systems,omitempty"`
Alerts []AlertInfo `json:"triggered_alerts,omitempty"`
Version string `json:"beszel_version"`
}
// SystemsSummary contains counts of systems by status.
type SystemsSummary struct {
Total int `json:"total"`
Up int `json:"up"`
Down int `json:"down"`
Paused int `json:"paused"`
Pending int `json:"pending"`
}
// SystemInfo identifies a system that is currently down.
type SystemInfo struct {
ID string `json:"id" db:"id"`
Name string `json:"name" db:"name"`
Host string `json:"host" db:"host"`
}
// AlertInfo describes a currently triggered alert.
type AlertInfo struct {
SystemID string `json:"system_id"`
SystemName string `json:"system_name"`
AlertName string `json:"alert_name"`
Threshold float64 `json:"threshold"`
}
// Config holds heartbeat settings read from environment variables.
type Config struct {
URL string // endpoint to ping
Interval int // seconds between pings
Method string // HTTP method (GET or POST, default POST)
}
// Heartbeat manages the periodic outbound health check.
type Heartbeat struct {
app core.App
config Config
client *http.Client
}
// New creates a Heartbeat if configuration is present.
// Returns nil if HEARTBEAT_URL is not set (feature disabled).
func New(app core.App, getEnv func(string) (string, bool)) *Heartbeat {
url, _ := getEnv("HEARTBEAT_URL")
url = strings.TrimSpace(url)
if app == nil || url == "" {
return nil
}
interval := defaultInterval
if v, ok := getEnv("HEARTBEAT_INTERVAL"); ok {
if parsed, err := strconv.Atoi(v); err == nil && parsed > 0 {
interval = parsed
}
}
method := http.MethodPost
if v, ok := getEnv("HEARTBEAT_METHOD"); ok {
v = strings.ToUpper(strings.TrimSpace(v))
if v == http.MethodGet || v == http.MethodHead {
method = v
}
}
return &Heartbeat{
app: app,
config: Config{
URL: url,
Interval: interval,
Method: method,
},
client: &http.Client{Timeout: httpTimeout},
}
}
// Start begins the heartbeat loop. It blocks and should be called in a goroutine.
// The loop runs until the provided stop channel is closed.
func (hb *Heartbeat) Start(stop <-chan struct{}) {
sanitizedURL := sanitizeHeartbeatURL(hb.config.URL)
hb.app.Logger().Info("Heartbeat enabled",
"url", sanitizedURL,
"interval", fmt.Sprintf("%ds", hb.config.Interval),
"method", hb.config.Method,
)
// Send an initial heartbeat immediately on startup.
hb.send()
ticker := time.NewTicker(time.Duration(hb.config.Interval) * time.Second)
defer ticker.Stop()
for {
select {
case <-stop:
return
case <-ticker.C:
hb.send()
}
}
}
// Send performs a single heartbeat ping. Exposed for the test-heartbeat API endpoint.
func (hb *Heartbeat) Send() error {
return hb.send()
}
// GetConfig returns the current heartbeat configuration.
func (hb *Heartbeat) GetConfig() Config {
return hb.config
}
func (hb *Heartbeat) send() error {
var req *http.Request
var err error
method := normalizeMethod(hb.config.Method)
if method == http.MethodGet || method == http.MethodHead {
req, err = http.NewRequest(method, hb.config.URL, nil)
} else {
payload, payloadErr := hb.buildPayload()
if payloadErr != nil {
hb.app.Logger().Error("Heartbeat: failed to build payload", "err", payloadErr)
return payloadErr
}
body, jsonErr := json.Marshal(payload)
if jsonErr != nil {
hb.app.Logger().Error("Heartbeat: failed to marshal payload", "err", jsonErr)
return jsonErr
}
req, err = http.NewRequest(http.MethodPost, hb.config.URL, bytes.NewReader(body))
if err == nil {
req.Header.Set("Content-Type", "application/json")
}
}
if err != nil {
hb.app.Logger().Error("Heartbeat: failed to create request", "err", err)
return err
}
req.Header.Set("User-Agent", "Beszel-Heartbeat")
resp, err := hb.client.Do(req)
if err != nil {
hb.app.Logger().Error("Heartbeat: request failed", "url", sanitizeHeartbeatURL(hb.config.URL), "err", err)
return err
}
defer resp.Body.Close()
if resp.StatusCode >= 400 {
hb.app.Logger().Warn("Heartbeat: non-success response",
"url", sanitizeHeartbeatURL(hb.config.URL),
"status", resp.StatusCode,
)
return fmt.Errorf("heartbeat endpoint returned status %d", resp.StatusCode)
}
return nil
}
func (hb *Heartbeat) buildPayload() (*Payload, error) {
db := hb.app.DB()
// Count systems by status.
var systemCounts []struct {
Status string `db:"status"`
Count int `db:"cnt"`
}
err := db.NewQuery("SELECT status, COUNT(*) as cnt FROM systems GROUP BY status").All(&systemCounts)
if err != nil {
return nil, fmt.Errorf("query system counts: %w", err)
}
summary := SystemsSummary{}
for _, sc := range systemCounts {
switch sc.Status {
case "up":
summary.Up = sc.Count
case "down":
summary.Down = sc.Count
case "paused":
summary.Paused = sc.Count
case "pending":
summary.Pending = sc.Count
}
summary.Total += sc.Count
}
// Get names of down systems.
var downSystems []SystemInfo
if summary.Down > 0 {
err = db.NewQuery("SELECT id, name, host FROM systems WHERE status = 'down'").All(&downSystems)
if err != nil {
return nil, fmt.Errorf("query down systems: %w", err)
}
}
// Get triggered alerts with system names.
var triggeredAlerts []struct {
SystemID string `db:"system"`
SystemName string `db:"system_name"`
AlertName string `db:"name"`
Value float64 `db:"value"`
}
err = db.NewQuery(`
SELECT a.system, s.name as system_name, a.name, a.value
FROM alerts a
JOIN systems s ON a.system = s.id
WHERE a.triggered = true
`).All(&triggeredAlerts)
if err != nil {
// Non-fatal: alerts info is supplementary.
triggeredAlerts = nil
}
alerts := make([]AlertInfo, 0, len(triggeredAlerts))
for _, ta := range triggeredAlerts {
alerts = append(alerts, AlertInfo{
SystemID: ta.SystemID,
SystemName: ta.SystemName,
AlertName: ta.AlertName,
Threshold: ta.Value,
})
}
// Determine overall status.
status := "ok"
msg := "All systems operational"
if summary.Down > 0 {
status = "error"
names := make([]string, len(downSystems))
for i, ds := range downSystems {
names[i] = ds.Name
}
msg = fmt.Sprintf("%d system(s) down: %s", summary.Down, strings.Join(names, ", "))
} else if len(alerts) > 0 {
status = "warn"
msg = fmt.Sprintf("%d alert(s) triggered", len(alerts))
}
return &Payload{
Status: status,
Timestamp: time.Now().UTC().Format(time.RFC3339),
Msg: msg,
Systems: summary,
Down: downSystems,
Alerts: alerts,
Version: beszel.Version,
}, nil
}
func normalizeMethod(method string) string {
upper := strings.ToUpper(strings.TrimSpace(method))
if upper == http.MethodGet || upper == http.MethodHead || upper == http.MethodPost {
return upper
}
return http.MethodPost
}
func sanitizeHeartbeatURL(rawURL string) string {
parsed, err := url.Parse(strings.TrimSpace(rawURL))
if err != nil || parsed.Scheme == "" || parsed.Host == "" {
return "<invalid-url>"
}
return parsed.Scheme + "://" + parsed.Host
}

View File

@@ -1,257 +0,0 @@
//go:build testing
package heartbeat_test
import (
"encoding/json"
"io"
"net/http"
"net/http/httptest"
"testing"
"github.com/henrygd/beszel/internal/hub/heartbeat"
beszeltests "github.com/henrygd/beszel/internal/tests"
"github.com/pocketbase/pocketbase/core"
"github.com/stretchr/testify/assert"
"github.com/stretchr/testify/require"
)
func TestNew(t *testing.T) {
t.Run("returns nil when app is missing", func(t *testing.T) {
hb := heartbeat.New(nil, envGetter(map[string]string{
"HEARTBEAT_URL": "https://heartbeat.example.com/ping",
}))
assert.Nil(t, hb)
})
t.Run("returns nil when URL is missing", func(t *testing.T) {
app := newTestHub(t)
hb := heartbeat.New(app.App, func(string) (string, bool) {
return "", false
})
assert.Nil(t, hb)
})
t.Run("parses and normalizes config values", func(t *testing.T) {
app := newTestHub(t)
env := map[string]string{
"HEARTBEAT_URL": " https://heartbeat.example.com/ping ",
"HEARTBEAT_INTERVAL": "90",
"HEARTBEAT_METHOD": "head",
}
getEnv := func(key string) (string, bool) {
v, ok := env[key]
return v, ok
}
hb := heartbeat.New(app.App, getEnv)
require.NotNil(t, hb)
cfg := hb.GetConfig()
assert.Equal(t, "https://heartbeat.example.com/ping", cfg.URL)
assert.Equal(t, 90, cfg.Interval)
assert.Equal(t, http.MethodHead, cfg.Method)
})
}
func TestSendGETDoesNotRequireAppOrDB(t *testing.T) {
app := newTestHub(t)
server := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
assert.Equal(t, http.MethodGet, r.Method)
assert.Equal(t, "Beszel-Heartbeat", r.Header.Get("User-Agent"))
w.WriteHeader(http.StatusOK)
}))
defer server.Close()
hb := heartbeat.New(app.App, envGetter(map[string]string{
"HEARTBEAT_URL": server.URL,
"HEARTBEAT_METHOD": "GET",
}))
require.NotNil(t, hb)
require.NoError(t, hb.Send())
}
func TestSendReturnsErrorOnHTTPFailureStatus(t *testing.T) {
app := newTestHub(t)
server := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
w.WriteHeader(http.StatusInternalServerError)
}))
defer server.Close()
hb := heartbeat.New(app.App, envGetter(map[string]string{
"HEARTBEAT_URL": server.URL,
"HEARTBEAT_METHOD": "GET",
}))
require.NotNil(t, hb)
err := hb.Send()
require.Error(t, err)
assert.ErrorContains(t, err, "heartbeat endpoint returned status 500")
}
func TestSendPOSTBuildsExpectedStatuses(t *testing.T) {
tests := []struct {
name string
setup func(t *testing.T, app *beszeltests.TestHub, user *core.Record)
expectStatus string
expectMsgPart string
expectDown int
expectAlerts int
expectTotal int
expectUp int
expectPaused int
expectPending int
expectDownSumm int
}{
{
name: "error when at least one system is down",
setup: func(t *testing.T, app *beszeltests.TestHub, user *core.Record) {
downSystem := createTestSystem(t, app, user.Id, "db-1", "10.0.0.1", "down")
_ = createTestSystem(t, app, user.Id, "web-1", "10.0.0.2", "up")
createTriggeredAlert(t, app, user.Id, downSystem.Id, "CPU", 95)
},
expectStatus: "error",
expectMsgPart: "1 system(s) down",
expectDown: 1,
expectAlerts: 1,
expectTotal: 2,
expectUp: 1,
expectDownSumm: 1,
},
{
name: "warn when only alerts are triggered",
setup: func(t *testing.T, app *beszeltests.TestHub, user *core.Record) {
system := createTestSystem(t, app, user.Id, "api-1", "10.1.0.1", "up")
createTriggeredAlert(t, app, user.Id, system.Id, "CPU", 90)
},
expectStatus: "warn",
expectMsgPart: "1 alert(s) triggered",
expectDown: 0,
expectAlerts: 1,
expectTotal: 1,
expectUp: 1,
expectDownSumm: 0,
},
{
name: "ok when no down systems and no alerts",
setup: func(t *testing.T, app *beszeltests.TestHub, user *core.Record) {
_ = createTestSystem(t, app, user.Id, "node-1", "10.2.0.1", "up")
_ = createTestSystem(t, app, user.Id, "node-2", "10.2.0.2", "paused")
_ = createTestSystem(t, app, user.Id, "node-3", "10.2.0.3", "pending")
},
expectStatus: "ok",
expectMsgPart: "All systems operational",
expectDown: 0,
expectAlerts: 0,
expectTotal: 3,
expectUp: 1,
expectPaused: 1,
expectPending: 1,
expectDownSumm: 0,
},
}
for _, tt := range tests {
t.Run(tt.name, func(t *testing.T) {
app := newTestHub(t)
user := createTestUser(t, app)
tt.setup(t, app, user)
type requestCapture struct {
method string
userAgent string
contentType string
payload heartbeat.Payload
}
captured := make(chan requestCapture, 1)
server := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
defer r.Body.Close()
body, err := io.ReadAll(r.Body)
require.NoError(t, err)
var payload heartbeat.Payload
require.NoError(t, json.Unmarshal(body, &payload))
captured <- requestCapture{
method: r.Method,
userAgent: r.Header.Get("User-Agent"),
contentType: r.Header.Get("Content-Type"),
payload: payload,
}
w.WriteHeader(http.StatusNoContent)
}))
defer server.Close()
hb := heartbeat.New(app.App, envGetter(map[string]string{
"HEARTBEAT_URL": server.URL,
"HEARTBEAT_METHOD": "POST",
}))
require.NotNil(t, hb)
require.NoError(t, hb.Send())
req := <-captured
assert.Equal(t, http.MethodPost, req.method)
assert.Equal(t, "Beszel-Heartbeat", req.userAgent)
assert.Equal(t, "application/json", req.contentType)
assert.Equal(t, tt.expectStatus, req.payload.Status)
assert.Contains(t, req.payload.Msg, tt.expectMsgPart)
assert.Equal(t, tt.expectDown, len(req.payload.Down))
assert.Equal(t, tt.expectAlerts, len(req.payload.Alerts))
assert.Equal(t, tt.expectTotal, req.payload.Systems.Total)
assert.Equal(t, tt.expectUp, req.payload.Systems.Up)
assert.Equal(t, tt.expectDownSumm, req.payload.Systems.Down)
assert.Equal(t, tt.expectPaused, req.payload.Systems.Paused)
assert.Equal(t, tt.expectPending, req.payload.Systems.Pending)
})
}
}
func newTestHub(t *testing.T) *beszeltests.TestHub {
t.Helper()
app, err := beszeltests.NewTestHub(t.TempDir())
require.NoError(t, err)
t.Cleanup(app.Cleanup)
return app
}
func createTestUser(t *testing.T, app *beszeltests.TestHub) *core.Record {
t.Helper()
user, err := beszeltests.CreateUser(app.App, "admin@example.com", "password123")
require.NoError(t, err)
return user
}
func createTestSystem(t *testing.T, app *beszeltests.TestHub, userID, name, host, status string) *core.Record {
t.Helper()
system, err := beszeltests.CreateRecord(app.App, "systems", map[string]any{
"name": name,
"host": host,
"port": "45876",
"users": []string{userID},
"status": status,
})
require.NoError(t, err)
return system
}
func createTriggeredAlert(t *testing.T, app *beszeltests.TestHub, userID, systemID, name string, threshold float64) *core.Record {
t.Helper()
alert, err := beszeltests.CreateRecord(app.App, "alerts", map[string]any{
"name": name,
"system": systemID,
"user": userID,
"value": threshold,
"min": 0,
"triggered": true,
})
require.NoError(t, err)
return alert
}
func envGetter(values map[string]string) func(string) (string, bool) {
return func(key string) (string, bool) {
v, ok := values[key]
return v, ok
}
}

View File

@@ -9,20 +9,17 @@ import (
"net/url"
"os"
"path"
"regexp"
"strings"
"time"
"github.com/henrygd/beszel"
"github.com/henrygd/beszel/internal/alerts"
"github.com/henrygd/beszel/internal/hub/config"
"github.com/henrygd/beszel/internal/hub/heartbeat"
"github.com/henrygd/beszel/internal/hub/systems"
"github.com/henrygd/beszel/internal/records"
"github.com/henrygd/beszel/internal/users"
"github.com/google/uuid"
"github.com/pocketbase/dbx"
"github.com/pocketbase/pocketbase"
"github.com/pocketbase/pocketbase/apis"
"github.com/pocketbase/pocketbase/core"
@@ -35,15 +32,11 @@ type Hub struct {
um *users.UserManager
rm *records.RecordManager
sm *systems.SystemManager
hb *heartbeat.Heartbeat
hbStop chan struct{}
pubKey string
signer ssh.Signer
appURL string
}
var containerIDPattern = regexp.MustCompile(`^[a-fA-F0-9]{12,64}$`)
// NewHub creates a new Hub instance with default configuration
func NewHub(app core.App) *Hub {
hub := &Hub{}
@@ -54,10 +47,6 @@ func NewHub(app core.App) *Hub {
hub.rm = records.NewRecordManager(hub)
hub.sm = systems.NewSystemManager(hub)
hub.appURL, _ = GetEnv("APP_URL")
hub.hb = heartbeat.New(app, GetEnv)
if hub.hb != nil {
hub.hbStop = make(chan struct{})
}
return hub
}
@@ -98,10 +87,6 @@ func (h *Hub) StartHub() error {
if err := h.sm.Initialize(); err != nil {
return err
}
// start heartbeat if configured
if h.hb != nil {
go h.hb.Start(h.hbStop)
}
return e.Next()
})
@@ -208,34 +193,7 @@ func setCollectionAuthSettings(app core.App) error {
}
containersListRule := strings.Replace(systemsReadRule, "users.id", "system.users.id", 1)
containersCollection.ListRule = &containersListRule
if err := app.Save(containersCollection); err != nil {
return err
}
// allow all users to access system-related collections if SHARE_ALL_SYSTEMS is set
// these collections all have a "system" relation field
systemRelatedCollections := []string{"system_details", "smart_devices", "systemd_services"}
for _, collectionName := range systemRelatedCollections {
collection, err := app.FindCollectionByNameOrId(collectionName)
if err != nil {
return err
}
collection.ListRule = &containersListRule
// set viewRule for collections that need it (system_details, smart_devices)
if collection.ViewRule != nil {
collection.ViewRule = &containersListRule
}
// set deleteRule for smart_devices (allows user to dismiss disk warnings)
if collectionName == "smart_devices" {
deleteRule := containersListRule + " && @request.auth.role != \"readonly\""
collection.DeleteRule = &deleteRule
}
if err := app.Save(collection); err != nil {
return err
}
}
return nil
return app.Save(containersCollection)
}
// registerCronJobs sets up scheduled tasks
@@ -301,9 +259,6 @@ func (h *Hub) registerApiRoutes(se *core.ServeEvent) error {
})
// send test notification
apiAuth.POST("/test-notification", h.SendTestNotification)
// heartbeat status and test
apiAuth.GET("/heartbeat-status", h.getHeartbeatStatus)
apiAuth.POST("/test-heartbeat", h.testHeartbeat)
// get config.yml content
apiAuth.GET("/config-yaml", config.GetYamlConfig)
// handle agent websocket connection
@@ -313,8 +268,8 @@ func (h *Hub) registerApiRoutes(se *core.ServeEvent) error {
// update / delete user alerts
apiAuth.POST("/user-alerts", alerts.UpsertUserAlerts)
apiAuth.DELETE("/user-alerts", alerts.DeleteUserAlerts)
// refresh SMART devices for a system
apiAuth.POST("/smart/refresh", h.refreshSmartData)
// get SMART data
apiAuth.GET("/smart", h.getSmartData)
// get systemd service details
apiAuth.GET("/systemd/info", h.getSystemdInfo)
// /containers routes
@@ -333,129 +288,27 @@ func (h *Hub) getUniversalToken(e *core.RequestEvent) error {
userID := e.Auth.Id
query := e.Request.URL.Query()
token := query.Get("token")
enable := query.Get("enable")
permanent := query.Get("permanent")
// helper for deleting any existing permanent token record for this user
deletePermanent := func() error {
rec, err := h.FindFirstRecordByFilter("universal_tokens", "user = {:user}", dbx.Params{"user": userID})
if err != nil {
return nil // no record
}
return h.Delete(rec)
}
// helper for upserting a permanent token record for this user
upsertPermanent := func(token string) error {
rec, err := h.FindFirstRecordByFilter("universal_tokens", "user = {:user}", dbx.Params{"user": userID})
if err == nil {
rec.Set("token", token)
return h.Save(rec)
}
col, err := h.FindCachedCollectionByNameOrId("universal_tokens")
if err != nil {
return err
}
newRec := core.NewRecord(col)
newRec.Set("user", userID)
newRec.Set("token", token)
return h.Save(newRec)
}
// Disable universal tokens (both ephemeral and permanent)
if enable == "0" {
tokenMap.RemovebyValue(userID)
_ = deletePermanent()
return e.JSON(http.StatusOK, map[string]any{"token": token, "active": false, "permanent": false})
}
// Enable universal token (ephemeral or permanent)
if enable == "1" {
if token == "" {
token = uuid.New().String()
}
if permanent == "1" {
// make token permanent (persist across restarts)
tokenMap.RemovebyValue(userID)
if err := upsertPermanent(token); err != nil {
return err
}
return e.JSON(http.StatusOK, map[string]any{"token": token, "active": true, "permanent": true})
}
// default: ephemeral mode (1 hour)
_ = deletePermanent()
tokenMap.Set(token, userID, time.Hour)
return e.JSON(http.StatusOK, map[string]any{"token": token, "active": true, "permanent": false})
}
// Read current state
// Prefer permanent token if it exists.
if rec, err := h.FindFirstRecordByFilter("universal_tokens", "user = {:user}", dbx.Params{"user": userID}); err == nil {
dbToken := rec.GetString("token")
// If no token was provided, or the caller is asking about their permanent token, return it.
if token == "" || token == dbToken {
return e.JSON(http.StatusOK, map[string]any{"token": dbToken, "active": true, "permanent": true})
}
// Token doesn't match their permanent token (avoid leaking other info)
return e.JSON(http.StatusOK, map[string]any{"token": token, "active": false, "permanent": false})
}
// No permanent token; fall back to ephemeral token map.
if token == "" {
// return existing token if it exists
if token, _, ok := tokenMap.GetByValue(userID); ok {
return e.JSON(http.StatusOK, map[string]any{"token": token, "active": true, "permanent": false})
return e.JSON(http.StatusOK, map[string]any{"token": token, "active": true})
}
// if no token is provided, generate a new one
token = uuid.New().String()
}
response := map[string]any{"token": token}
// Token is considered active only if it belongs to the current user.
activeUser, ok := tokenMap.GetOk(token)
active := ok && activeUser == userID
response := map[string]any{"token": token, "active": active, "permanent": false}
switch query.Get("enable") {
case "1":
tokenMap.Set(token, userID, time.Hour)
case "0":
tokenMap.RemovebyValue(userID)
}
_, response["active"] = tokenMap.GetOk(token)
return e.JSON(http.StatusOK, response)
}
// getHeartbeatStatus returns current heartbeat configuration and whether it's enabled
func (h *Hub) getHeartbeatStatus(e *core.RequestEvent) error {
if e.Auth.GetString("role") != "admin" {
return e.ForbiddenError("Requires admin role", nil)
}
if h.hb == nil {
return e.JSON(http.StatusOK, map[string]any{
"enabled": false,
"msg": "Set HEARTBEAT_URL to enable outbound heartbeat monitoring",
})
}
cfg := h.hb.GetConfig()
return e.JSON(http.StatusOK, map[string]any{
"enabled": true,
"url": cfg.URL,
"interval": cfg.Interval,
"method": cfg.Method,
})
}
// testHeartbeat triggers a single heartbeat ping and returns the result
func (h *Hub) testHeartbeat(e *core.RequestEvent) error {
if e.Auth.GetString("role") != "admin" {
return e.ForbiddenError("Requires admin role", nil)
}
if h.hb == nil {
return e.JSON(http.StatusOK, map[string]any{
"err": "Heartbeat not configured. Set HEARTBEAT_URL environment variable.",
})
}
if err := h.hb.Send(); err != nil {
return e.JSON(http.StatusOK, map[string]any{"err": err.Error()})
}
return e.JSON(http.StatusOK, map[string]any{"err": false})
}
// containerRequestHandler handles both container logs and info requests
func (h *Hub) containerRequestHandler(e *core.RequestEvent, fetchFunc func(*systems.System, string) (string, error), responseKey string) error {
systemID := e.Request.URL.Query().Get("system")
@@ -464,9 +317,6 @@ func (h *Hub) containerRequestHandler(e *core.RequestEvent, fetchFunc func(*syst
if systemID == "" || containerID == "" {
return e.JSON(http.StatusBadRequest, map[string]string{"error": "system and container parameters are required"})
}
if !containerIDPattern.MatchString(containerID) {
return e.JSON(http.StatusBadRequest, map[string]string{"error": "invalid container parameter"})
}
system, err := h.sm.GetSystem(systemID)
if err != nil {
@@ -515,25 +365,22 @@ func (h *Hub) getSystemdInfo(e *core.RequestEvent) error {
return e.JSON(http.StatusOK, map[string]any{"details": details})
}
// refreshSmartData handles POST /api/beszel/smart/refresh requests
// Fetches fresh SMART data from the agent and updates the collection
func (h *Hub) refreshSmartData(e *core.RequestEvent) error {
// getSmartData handles GET /api/beszel/smart requests
func (h *Hub) getSmartData(e *core.RequestEvent) error {
systemID := e.Request.URL.Query().Get("system")
if systemID == "" {
return e.JSON(http.StatusBadRequest, map[string]string{"error": "system parameter is required"})
}
system, err := h.sm.GetSystem(systemID)
if err != nil {
return e.JSON(http.StatusNotFound, map[string]string{"error": "system not found"})
}
// Fetch and save SMART devices
if err := system.FetchAndSaveSmartDevices(); err != nil {
return e.JSON(http.StatusInternalServerError, map[string]string{"error": err.Error()})
data, err := system.FetchSmartDataFromAgent()
if err != nil {
return e.JSON(http.StatusNotFound, map[string]string{"error": err.Error()})
}
return e.JSON(http.StatusOK, map[string]string{"status": "ok"})
e.Response.Header().Set("Cache-Control", "public, max-age=60")
return e.JSON(http.StatusOK, data)
}
// generates key pair if it doesn't exist and returns signer

View File

@@ -1,4 +1,5 @@
//go:build testing
// +build testing
package hub_test
@@ -361,58 +362,6 @@ func TestApiRoutesAuthentication(t *testing.T) {
ExpectedContent: []string{"test-system"},
TestAppFactory: testAppFactory,
},
{
Name: "GET /heartbeat-status - no auth should fail",
Method: http.MethodGet,
URL: "/api/beszel/heartbeat-status",
ExpectedStatus: 401,
ExpectedContent: []string{"requires valid"},
TestAppFactory: testAppFactory,
},
{
Name: "GET /heartbeat-status - with user auth should fail",
Method: http.MethodGet,
URL: "/api/beszel/heartbeat-status",
Headers: map[string]string{
"Authorization": userToken,
},
ExpectedStatus: 403,
ExpectedContent: []string{"Requires admin role"},
TestAppFactory: testAppFactory,
},
{
Name: "GET /heartbeat-status - with admin auth should succeed",
Method: http.MethodGet,
URL: "/api/beszel/heartbeat-status",
Headers: map[string]string{
"Authorization": adminUserToken,
},
ExpectedStatus: 200,
ExpectedContent: []string{`"enabled":false`},
TestAppFactory: testAppFactory,
},
{
Name: "POST /test-heartbeat - with user auth should fail",
Method: http.MethodPost,
URL: "/api/beszel/test-heartbeat",
Headers: map[string]string{
"Authorization": userToken,
},
ExpectedStatus: 403,
ExpectedContent: []string{"Requires admin role"},
TestAppFactory: testAppFactory,
},
{
Name: "POST /test-heartbeat - with admin auth should report disabled state",
Method: http.MethodPost,
URL: "/api/beszel/test-heartbeat",
Headers: map[string]string{
"Authorization": adminUserToken,
},
ExpectedStatus: 200,
ExpectedContent: []string{"Heartbeat not configured"},
TestAppFactory: testAppFactory,
},
{
Name: "GET /universal-token - no auth should fail",
Method: http.MethodGet,
@@ -429,18 +378,7 @@ func TestApiRoutesAuthentication(t *testing.T) {
"Authorization": userToken,
},
ExpectedStatus: 200,
ExpectedContent: []string{"active", "token", "permanent"},
TestAppFactory: testAppFactory,
},
{
Name: "GET /universal-token - enable permanent should succeed",
Method: http.MethodGet,
URL: "/api/beszel/universal-token?enable=1&permanent=1&token=permanent-token-123",
Headers: map[string]string{
"Authorization": userToken,
},
ExpectedStatus: 200,
ExpectedContent: []string{"\"permanent\":true", "permanent-token-123"},
ExpectedContent: []string{"active", "token"},
TestAppFactory: testAppFactory,
},
{
@@ -544,7 +482,7 @@ func TestApiRoutesAuthentication(t *testing.T) {
{
Name: "GET /containers/logs - with auth but invalid system should fail",
Method: http.MethodGet,
URL: "/api/beszel/containers/logs?system=invalid-system&container=0123456789ab",
URL: "/api/beszel/containers/logs?system=invalid-system&container=test-container",
Headers: map[string]string{
"Authorization": userToken,
},
@@ -552,39 +490,6 @@ func TestApiRoutesAuthentication(t *testing.T) {
ExpectedContent: []string{"system not found"},
TestAppFactory: testAppFactory,
},
{
Name: "GET /containers/logs - traversal container should fail validation",
Method: http.MethodGet,
URL: "/api/beszel/containers/logs?system=" + system.Id + "&container=..%2F..%2Fversion",
Headers: map[string]string{
"Authorization": userToken,
},
ExpectedStatus: 400,
ExpectedContent: []string{"invalid container parameter"},
TestAppFactory: testAppFactory,
},
{
Name: "GET /containers/info - traversal container should fail validation",
Method: http.MethodGet,
URL: "/api/beszel/containers/info?system=" + system.Id + "&container=../../version?x=",
Headers: map[string]string{
"Authorization": userToken,
},
ExpectedStatus: 400,
ExpectedContent: []string{"invalid container parameter"},
TestAppFactory: testAppFactory,
},
{
Name: "GET /containers/info - non-hex container should fail validation",
Method: http.MethodGet,
URL: "/api/beszel/containers/info?system=" + system.Id + "&container=container_name",
Headers: map[string]string{
"Authorization": userToken,
},
ExpectedStatus: 400,
ExpectedContent: []string{"invalid container parameter"},
TestAppFactory: testAppFactory,
},
// Auth Optional Routes - Should work without authentication
{

View File

@@ -1,4 +1,5 @@
//go:build testing
// +build testing
package hub

View File

@@ -9,15 +9,12 @@ import (
"math/rand"
"net"
"strings"
"sync/atomic"
"time"
"github.com/henrygd/beszel/internal/common"
"github.com/henrygd/beszel/internal/hub/transport"
"github.com/henrygd/beszel/internal/hub/ws"
"github.com/henrygd/beszel/internal/entities/container"
"github.com/henrygd/beszel/internal/entities/smart"
"github.com/henrygd/beszel/internal/entities/system"
"github.com/henrygd/beszel/internal/entities/systemd"
@@ -25,30 +22,24 @@ import (
"github.com/blang/semver"
"github.com/fxamacker/cbor/v2"
"github.com/lxzan/gws"
"github.com/pocketbase/dbx"
"github.com/pocketbase/pocketbase/core"
"golang.org/x/crypto/ssh"
)
type System struct {
Id string `db:"id"`
Host string `db:"host"`
Port string `db:"port"`
Status string `db:"status"`
manager *SystemManager // Manager that this system belongs to
client *ssh.Client // SSH client for fetching data
sshTransport *transport.SSHTransport // SSH transport for requests
data *system.CombinedData // system data from agent
ctx context.Context // Context for stopping the updater
cancel context.CancelFunc // Stops and removes system from updater
WsConn *ws.WsConn // Handler for agent WebSocket connection
agentVersion semver.Version // Agent version
updateTicker *time.Ticker // Ticker for updating the system
detailsFetched atomic.Bool // True if static system details have been fetched and saved
smartFetching atomic.Bool // True if SMART devices are currently being fetched
smartInterval time.Duration // Interval for periodic SMART data updates
lastSmartFetch atomic.Int64 // Unix milliseconds of last SMART data fetch
Id string `db:"id"`
Host string `db:"host"`
Port string `db:"port"`
Status string `db:"status"`
manager *SystemManager // Manager that this system belongs to
client *ssh.Client // SSH client for fetching data
data *system.CombinedData // system data from agent
ctx context.Context // Context for stopping the updater
cancel context.CancelFunc // Stops and removes system from updater
WsConn *ws.WsConn // Handler for agent WebSocket connection
agentVersion semver.Version // Agent version
updateTicker *time.Ticker // Ticker for updating the system
}
func (sm *SystemManager) NewSystem(systemId string) *System {
@@ -121,37 +112,10 @@ func (sys *System) update() error {
sys.handlePaused()
return nil
}
options := common.DataRequestOptions{
CacheTimeMs: uint16(interval),
data, err := sys.fetchDataFromAgent(common.DataRequestOptions{CacheTimeMs: uint16(interval)})
if err == nil {
_, err = sys.createRecords(data)
}
// fetch system details if not already fetched
if !sys.detailsFetched.Load() {
options.IncludeDetails = true
}
data, err := sys.fetchDataFromAgent(options)
if err != nil {
return err
}
// create system records
_, err = sys.createRecords(data)
// Fetch and save SMART devices when system first comes online or at intervals
if backgroundSmartFetchEnabled() {
if sys.smartInterval <= 0 {
sys.smartInterval = time.Hour
}
lastFetch := sys.lastSmartFetch.Load()
if time.Since(time.UnixMilli(lastFetch)) >= sys.smartInterval && sys.smartFetching.CompareAndSwap(false, true) {
go func() {
defer sys.smartFetching.Store(false)
sys.lastSmartFetch.Store(time.Now().UnixMilli())
_ = sys.FetchAndSaveSmartDevices()
}()
}
}
return err
}
@@ -176,11 +140,12 @@ func (sys *System) createRecords(data *system.CombinedData) (*core.Record, error
}
hub := sys.manager.hub
err = hub.RunInTransaction(func(txApp core.App) error {
// add system_stats record
// add system_stats and container_stats records
systemStatsCollection, err := txApp.FindCachedCollectionByNameOrId("system_stats")
if err != nil {
return err
}
systemStatsRecord := core.NewRecord(systemStatsCollection)
systemStatsRecord.Set("system", systemRecord.Id)
systemStatsRecord.Set("stats", data.Stats)
@@ -188,14 +153,14 @@ func (sys *System) createRecords(data *system.CombinedData) (*core.Record, error
if err := txApp.SaveNoValidate(systemStatsRecord); err != nil {
return err
}
// add containers and container_stats records
if len(data.Containers) > 0 {
// add / update containers records
if data.Containers[0].Id != "" {
if err := createContainerRecords(txApp, data.Containers, sys.Id); err != nil {
return err
}
}
// add new container_stats record
containerStatsCollection, err := txApp.FindCachedCollectionByNameOrId("container_stats")
if err != nil {
return err
@@ -216,20 +181,9 @@ func (sys *System) createRecords(data *system.CombinedData) (*core.Record, error
}
}
// add system details record
if data.Details != nil {
if err := createSystemDetailsRecord(txApp, data.Details, sys.Id); err != nil {
return err
}
sys.detailsFetched.Store(true)
// update smart interval if it's set on the agent side
if data.Details.SmartInterval > 0 {
sys.smartInterval = data.Details.SmartInterval
}
}
// update system record (do this last because it triggers alerts and we need above records to be inserted first)
systemRecord.Set("status", up)
systemRecord.Set("info", data.Info)
if err := txApp.SaveNoValidate(systemRecord); err != nil {
return err
@@ -240,31 +194,6 @@ func (sys *System) createRecords(data *system.CombinedData) (*core.Record, error
return systemRecord, err
}
func createSystemDetailsRecord(app core.App, data *system.Details, systemId string) error {
collectionName := "system_details"
params := dbx.Params{
"id": systemId,
"system": systemId,
"hostname": data.Hostname,
"kernel": data.Kernel,
"cores": data.Cores,
"threads": data.Threads,
"cpu": data.CpuModel,
"os": data.Os,
"os_name": data.OsName,
"arch": data.Arch,
"memory": data.MemoryTotal,
"podman": data.Podman,
"updated": time.Now().UTC(),
}
result, err := app.DB().Update(collectionName, params, dbx.HashExp{"id": systemId}).Execute()
rowsAffected, _ := result.RowsAffected()
if err != nil || rowsAffected == 0 {
_, err = app.DB().Insert(collectionName, params).Execute()
}
return err
}
func createSystemdStatsRecords(app core.App, data []*systemd.Service, systemId string) error {
if len(data) == 0 {
return nil
@@ -279,7 +208,7 @@ func createSystemdStatsRecords(app core.App, data []*systemd.Service, systemId s
for i, service := range data {
suffix := fmt.Sprintf("%d", i)
valueStrings = append(valueStrings, fmt.Sprintf("({:id%[1]s}, {:system}, {:name%[1]s}, {:state%[1]s}, {:sub%[1]s}, {:cpu%[1]s}, {:cpuPeak%[1]s}, {:memory%[1]s}, {:memPeak%[1]s}, {:updated})", suffix))
params["id"+suffix] = makeStableHashId(systemId, service.Name)
params["id"+suffix] = getSystemdServiceId(systemId, service.Name)
params["name"+suffix] = service.Name
params["state"+suffix] = service.State
params["sub"+suffix] = service.Sub
@@ -296,6 +225,13 @@ func createSystemdStatsRecords(app core.App, data []*systemd.Service, systemId s
return err
}
// getSystemdServiceId generates a deterministic unique id for a systemd service
func getSystemdServiceId(systemId string, serviceName string) string {
hash := fnv.New32a()
hash.Write([]byte(systemId + serviceName))
return fmt.Sprintf("%x", hash.Sum32())
}
// createContainerRecords creates container records
func createContainerRecords(app core.App, data []*container.Stats, systemId string) error {
if len(data) == 0 {
@@ -317,11 +253,7 @@ func createContainerRecords(app core.App, data []*container.Stats, systemId stri
params["health"+suffix] = container.Health
params["cpu"+suffix] = container.Cpu
params["memory"+suffix] = container.Mem
netBytes := container.Bandwidth[0] + container.Bandwidth[1]
if netBytes == 0 {
netBytes = uint64((container.NetworkSent + container.NetworkRecv) * 1024 * 1024)
}
params["net"+suffix] = netBytes
params["net"+suffix] = container.NetworkSent + container.NetworkRecv
}
queryString := fmt.Sprintf(
"INSERT INTO containers (id, system, name, image, status, health, cpu, memory, net, updated) VALUES %s ON CONFLICT(id) DO UPDATE SET system = excluded.system, name = excluded.name, image = excluded.image, status = excluded.status, health = excluded.health, cpu = excluded.cpu, memory = excluded.memory, net = excluded.net, updated = excluded.updated",
@@ -367,78 +299,8 @@ func (sys *System) getContext() (context.Context, context.CancelFunc) {
return sys.ctx, sys.cancel
}
// request sends a request to the agent, trying WebSocket first, then SSH.
// This is the unified request method that uses the transport abstraction.
func (sys *System) request(ctx context.Context, action common.WebSocketAction, req any, dest any) error {
// Try WebSocket first
if sys.WsConn != nil && sys.WsConn.IsConnected() {
wsTransport := transport.NewWebSocketTransport(sys.WsConn)
if err := wsTransport.Request(ctx, action, req, dest); err == nil {
return nil
} else if !shouldFallbackToSSH(err) {
return err
} else if shouldCloseWebSocket(err) {
sys.closeWebSocketConnection()
}
}
// Fall back to SSH if WebSocket fails
if err := sys.ensureSSHTransport(); err != nil {
return err
}
err := sys.sshTransport.RequestWithRetry(ctx, action, req, dest, 1)
// Keep legacy SSH client/version fields in sync for other code paths.
if sys.sshTransport != nil {
sys.client = sys.sshTransport.GetClient()
sys.agentVersion = sys.sshTransport.GetAgentVersion()
}
return err
}
func shouldFallbackToSSH(err error) bool {
if err == nil {
return false
}
if errors.Is(err, context.DeadlineExceeded) || errors.Is(err, context.Canceled) {
return true
}
if errors.Is(err, gws.ErrConnClosed) {
return true
}
return errors.Is(err, transport.ErrWebSocketNotConnected)
}
func shouldCloseWebSocket(err error) bool {
if err == nil {
return false
}
return errors.Is(err, gws.ErrConnClosed) || errors.Is(err, transport.ErrWebSocketNotConnected)
}
// ensureSSHTransport ensures the SSH transport is initialized and connected.
func (sys *System) ensureSSHTransport() error {
if sys.sshTransport == nil {
if sys.manager.sshConfig == nil {
if err := sys.manager.createSSHClientConfig(); err != nil {
return err
}
}
sys.sshTransport = transport.NewSSHTransport(transport.SSHTransportConfig{
Host: sys.Host,
Port: sys.Port,
Config: sys.manager.sshConfig,
Timeout: 4 * time.Second,
})
}
// Sync client state with transport
if sys.client != nil {
sys.sshTransport.SetClient(sys.client)
sys.sshTransport.SetAgentVersion(sys.agentVersion)
}
return nil
}
// fetchDataFromAgent attempts to fetch data from the agent, prioritizing WebSocket if available.
// fetchDataFromAgent attempts to fetch data from the agent,
// prioritizing WebSocket if available.
func (sys *System) fetchDataFromAgent(options common.DataRequestOptions) (*system.CombinedData, error) {
if sys.data == nil {
sys.data = &system.CombinedData{}
@@ -464,56 +326,152 @@ func (sys *System) fetchDataViaWebSocket(options common.DataRequestOptions) (*sy
if sys.WsConn == nil || !sys.WsConn.IsConnected() {
return nil, errors.New("no websocket connection")
}
wsTransport := transport.NewWebSocketTransport(sys.WsConn)
err := wsTransport.Request(context.Background(), common.GetData, options, sys.data)
err := sys.WsConn.RequestSystemData(context.Background(), sys.data, options)
if err != nil {
return nil, err
}
return sys.data, nil
}
// fetchStringFromAgentViaSSH is a generic function to fetch strings via SSH
func (sys *System) fetchStringFromAgentViaSSH(action common.WebSocketAction, requestData any, errorMsg string) (string, error) {
var result string
err := sys.runSSHOperation(4*time.Second, 1, func(session *ssh.Session) (bool, error) {
stdout, err := session.StdoutPipe()
if err != nil {
return false, err
}
stdin, stdinErr := session.StdinPipe()
if stdinErr != nil {
return false, stdinErr
}
if err := session.Shell(); err != nil {
return false, err
}
req := common.HubRequest[any]{Action: action, Data: requestData}
_ = cbor.NewEncoder(stdin).Encode(req)
_ = stdin.Close()
var resp common.AgentResponse
err = cbor.NewDecoder(stdout).Decode(&resp)
if err != nil {
return false, err
}
if resp.String == nil {
return false, errors.New(errorMsg)
}
result = *resp.String
return false, nil
})
return result, err
}
// FetchContainerInfoFromAgent fetches container info from the agent
func (sys *System) FetchContainerInfoFromAgent(containerID string) (string, error) {
ctx, cancel := context.WithTimeout(context.Background(), 5*time.Second)
defer cancel()
var result string
err := sys.request(ctx, common.GetContainerInfo, common.ContainerInfoRequest{ContainerID: containerID}, &result)
return result, err
// fetch via websocket
if sys.WsConn != nil && sys.WsConn.IsConnected() {
ctx, cancel := context.WithTimeout(context.Background(), 5*time.Second)
defer cancel()
return sys.WsConn.RequestContainerInfo(ctx, containerID)
}
// fetch via SSH
return sys.fetchStringFromAgentViaSSH(common.GetContainerInfo, common.ContainerInfoRequest{ContainerID: containerID}, "no info in response")
}
// FetchContainerLogsFromAgent fetches container logs from the agent
func (sys *System) FetchContainerLogsFromAgent(containerID string) (string, error) {
ctx, cancel := context.WithTimeout(context.Background(), 5*time.Second)
defer cancel()
var result string
err := sys.request(ctx, common.GetContainerLogs, common.ContainerLogsRequest{ContainerID: containerID}, &result)
return result, err
// fetch via websocket
if sys.WsConn != nil && sys.WsConn.IsConnected() {
ctx, cancel := context.WithTimeout(context.Background(), 5*time.Second)
defer cancel()
return sys.WsConn.RequestContainerLogs(ctx, containerID)
}
// fetch via SSH
return sys.fetchStringFromAgentViaSSH(common.GetContainerLogs, common.ContainerLogsRequest{ContainerID: containerID}, "no logs in response")
}
// FetchSystemdInfoFromAgent fetches detailed systemd service information from the agent
func (sys *System) FetchSystemdInfoFromAgent(serviceName string) (systemd.ServiceDetails, error) {
ctx, cancel := context.WithTimeout(context.Background(), 5*time.Second)
defer cancel()
// fetch via websocket
if sys.WsConn != nil && sys.WsConn.IsConnected() {
ctx, cancel := context.WithTimeout(context.Background(), 5*time.Second)
defer cancel()
return sys.WsConn.RequestSystemdInfo(ctx, serviceName)
}
var result systemd.ServiceDetails
err := sys.request(ctx, common.GetSystemdInfo, common.SystemdInfoRequest{ServiceName: serviceName}, &result)
err := sys.runSSHOperation(5*time.Second, 1, func(session *ssh.Session) (bool, error) {
stdout, err := session.StdoutPipe()
if err != nil {
return false, err
}
stdin, stdinErr := session.StdinPipe()
if stdinErr != nil {
return false, stdinErr
}
if err := session.Shell(); err != nil {
return false, err
}
req := common.HubRequest[any]{Action: common.GetSystemdInfo, Data: common.SystemdInfoRequest{ServiceName: serviceName}}
if err := cbor.NewEncoder(stdin).Encode(req); err != nil {
return false, err
}
_ = stdin.Close()
var resp common.AgentResponse
if err := cbor.NewDecoder(stdout).Decode(&resp); err != nil {
return false, err
}
if resp.ServiceInfo == nil {
if resp.Error != "" {
return false, errors.New(resp.Error)
}
return false, errors.New("no systemd info in response")
}
result = resp.ServiceInfo
return false, nil
})
return result, err
}
// FetchSmartDataFromAgent fetches SMART data from the agent
func (sys *System) FetchSmartDataFromAgent() (map[string]smart.SmartData, error) {
ctx, cancel := context.WithTimeout(context.Background(), 60*time.Second)
defer cancel()
var result map[string]smart.SmartData
err := sys.request(ctx, common.GetSmartData, nil, &result)
return result, err
}
func makeStableHashId(strings ...string) string {
hash := fnv.New32a()
for _, str := range strings {
hash.Write([]byte(str))
func (sys *System) FetchSmartDataFromAgent() (map[string]any, error) {
// fetch via websocket
if sys.WsConn != nil && sys.WsConn.IsConnected() {
ctx, cancel := context.WithTimeout(context.Background(), 5*time.Second)
defer cancel()
return sys.WsConn.RequestSmartData(ctx)
}
return fmt.Sprintf("%x", hash.Sum32())
// fetch via SSH
var result map[string]any
err := sys.runSSHOperation(5*time.Second, 1, func(session *ssh.Session) (bool, error) {
stdout, err := session.StdoutPipe()
if err != nil {
return false, err
}
stdin, stdinErr := session.StdinPipe()
if stdinErr != nil {
return false, stdinErr
}
if err := session.Shell(); err != nil {
return false, err
}
req := common.HubRequest[any]{Action: common.GetSmartData}
_ = cbor.NewEncoder(stdin).Encode(req)
_ = stdin.Close()
var resp common.AgentResponse
if err := cbor.NewDecoder(stdout).Decode(&resp); err != nil {
return false, err
}
// Convert to generic map for JSON response
result = make(map[string]any, len(resp.SmartData))
for k, v := range resp.SmartData {
result[k] = v
}
return false, nil
})
return result, err
}
// fetchDataViaSSH handles fetching data using SSH.
@@ -669,9 +627,6 @@ func (sys *System) createSessionWithTimeout(timeout time.Duration) (*ssh.Session
// closeSSHConnection closes the SSH connection but keeps the system in the manager
func (sys *System) closeSSHConnection() {
if sys.sshTransport != nil {
sys.sshTransport.Close()
}
if sys.client != nil {
sys.client.Close()
sys.client = nil

Some files were not shown because too many files have changed in this diff Show More