update

agent/agent.go

@@ -84,6 +84,7 @@ func NewAgent(dataDir ...string) (agent *Agent, err error) {
 			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) {
@@ -105,6 +106,16 @@ func NewAgent(dataDir ...string) (agent *Agent, err error) {
 	// 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)
+		}
+	}
+
 	// initialize connection manager
 	agent.connectionManager = newConnectionManager(agent)
 

agent/gpu.go

@@ -44,6 +44,7 @@ type GPUManager struct {
 	rocmSmi       bool
 	tegrastats    bool
 	intelGpuStats bool
+	nvml          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)
@@ -297,8 +298,13 @@ func (gm *GPUManager) calculateGPUAverage(id string, gpu *system.GPUData, cacheK
 	currentCount := uint32(gpu.Count)
 	deltaCount := gm.calculateDeltaCount(currentCount, lastSnapshot)
 
-	// If no new data arrived, use last known average
+	// If no new data arrived
 	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
 	}
 
@@ -396,7 +402,7 @@ func (gm *GPUManager) detectGPUs() error {
 	if _, err := exec.LookPath(intelGpuStatsCmd); err == nil {
 		gm.intelGpuStats = true
 	}
-	if gm.nvidiaSmi || gm.rocmSmi || gm.tegrastats || gm.intelGpuStats {
+	if gm.nvidiaSmi || gm.rocmSmi || gm.tegrastats || gm.intelGpuStats || gm.nvml {
 		return nil
 	}
 	return fmt.Errorf("no GPU found - install nvidia-smi, rocm-smi, tegrastats, or intel_gpu_top")
@@ -467,7 +473,20 @@ func NewGPUManager() (*GPUManager, error) {
 	gm.GpuDataMap = make(map[string]*system.GPUData)
 
 	if gm.nvidiaSmi {
-		gm.startCollector(nvidiaSmiCmd)
+		if nvml, _ := GetEnv("NVML"); nvml == "true" {
+			gm.nvml = true
+			gm.nvidiaSmi = false
+			collector := &nvmlCollector{gm: &gm}
+			if err := collector.init(); err == nil {
+				go collector.start()
+			} else {
+				slog.Warn("Failed to initialize NVML, falling back to nvidia-smi", "err", err)
+				gm.nvidiaSmi = true
+				gm.startCollector(nvidiaSmiCmd)
+			}
+		} else {
+			gm.startCollector(nvidiaSmiCmd)
+		}
 	}
 	if gm.rocmSmi {
 		gm.startCollector(rocmSmiCmd)

agent/gpu_nvml.go

@@ -0,0 +1,210 @@
+package agent
+
+import (
+	"fmt"
+	"strings"
+	"time"
+	"unsafe"
+
+	"github.com/ebitengine/purego"
+	"github.com/henrygd/beszel/internal/entities/system"
+	"golang.org/x/exp/slog"
+)
+
+// 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.Info("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.Info("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.Info("NVML: Utilization failed (GPU likely suspended)", "bdf", bdf, "ret", ret)
+			gpu.Temperature = 0
+			gpu.MemoryUsed = 0
+			continue
+		}
+
+		slog.Info("NVML: Collecting data for GPU", "bdf", bdf)
+
+		// Temperature
+		var temp uint32
+		nvmlDeviceGetTemperature(device, 0, &temp) // 0 is NVML_TEMPERATURE_GPU
+
+		// Memory
+		var usedMem, totalMem uint64
+		if c.isV2 {
+			var memory nvmlMemoryV2
+			memory.Version = 0x02000028 // (2 << 24) | 40 bytes
+			nvmlDeviceGetMemoryInfo(device, uintptr(unsafe.Pointer(&memory)))
+			usedMem = memory.Used
+			totalMem = memory.Total
+		} else {
+			var memory nvmlMemoryV1
+			nvmlDeviceGetMemoryInfo(device, uintptr(unsafe.Pointer(&memory)))
+			usedMem = memory.Used
+			totalMem = memory.Total
+		}
+
+		// Power
+		var power uint32
+		nvmlDeviceGetPowerUsage(device, &power)
+
+		gpu.Temperature = float64(temp)
+		gpu.MemoryUsed = float64(usedMem) / 1024 / 1024 / mebibytesInAMegabyte
+		gpu.MemoryTotal = float64(totalMem) / 1024 / 1024 / mebibytesInAMegabyte
+		gpu.Usage += float64(utilization.Gpu)
+		gpu.Power += float64(power) / 1000.0
+		gpu.Count++
+		slog.Info("NVML: Collected data", "gpu", gpu)
+	}
+}

agent/gpu_nvml_linux.go

@@ -0,0 +1,56 @@
+//go:build linux
+
+package agent
+
+import (
+	"os"
+	"path/filepath"
+	"strings"
+
+	"github.com/ebitengine/purego"
+	"golang.org/x/exp/slog"
+)
+
+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.Info("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.Info("NVML: GPU is 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" {
+				return false
+			}
+		}
+	}
+
+	return true
+}

agent/gpu_nvml_unsupported.go

@@ -0,0 +1,21 @@
+//go:build !linux && !windows
+
+package agent
+
+import "fmt"
+
+func openLibrary(name string) (uintptr, error) {
+	return 0, fmt.Errorf("nvml not supported on this platform")
+}
+
+func getNVMLPath() string {
+	return ""
+}
+
+func hasSymbol(lib uintptr, symbol string) bool {
+	return false
+}
+
+func (c *nvmlCollector) isGPUActive(bdf string) bool {
+	return true
+}

agent/gpu_nvml_windows.go

@@ -0,0 +1,25 @@
+//go:build windows
+
+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
+}

go.mod

@@ -6,6 +6,7 @@ require (
 	github.com/blang/semver v3.5.1+incompatible
 	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
@@ -20,6 +21,7 @@ require (
 	github.com/stretchr/testify v1.11.1
 	golang.org/x/crypto v0.45.0
 	golang.org/x/exp v0.0.0-20251125195548-87e1e737ad39
+	golang.org/x/sys v0.38.0
 	gopkg.in/yaml.v3 v3.0.1
 )
 
@@ -31,7 +33,6 @@ 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 v1.4.11 // indirect
 	github.com/ganigeorgiev/fexpr v0.5.0 // indirect
@@ -57,7 +58,6 @@ require (
 	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

go.sum

@@ -62,8 +62,6 @@ github.com/google/pprof v0.0.0-20251114195745-4902fdda35c8 h1:3DsUAV+VNEQa2CUVLx
 github.com/google/pprof v0.0.0-20251114195745-4902fdda35c8/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=
-github.com/hashicorp/golang-lru/v2 v2.0.7/go.mod h1:QeFd9opnmA6QUJc5vARoKUSoFhyfM2/ZepoAG6RGpeM=
 github.com/inconshreveable/mousetrap v1.1.0 h1:wN+x4NVGpMsO7ErUn/mUI3vEoE6Jt13X2s0bqwp9tc8=
 github.com/inconshreveable/mousetrap v1.1.0/go.mod h1:vpF70FUmC8bwa3OWnCshd2FqLfsEA9PFc4w1p2J65bw=
 github.com/jarcoal/httpmock v1.4.1 h1:0Ju+VCFuARfFlhVXFc2HxlcQkfB+Xq12/EotHko+x2A=
@@ -173,22 +171,18 @@ gopkg.in/yaml.v3 v3.0.1 h1:fxVm/GzAzEWqLHuvctI91KS9hhNmmWOoWu0XTYJS7CA=
 gopkg.in/yaml.v3 v3.0.1/go.mod h1:K4uyk7z7BCEPqu6E+C64Yfv1cQ7kz7rIZviUmN+EgEM=
 howett.net/plist v1.0.1 h1:37GdZ8tP09Q35o9ych3ehygcsL+HqKSwzctveSlarvM=
 howett.net/plist v1.0.1/go.mod h1:lqaXoTrLY4hg8tnEzNru53gicrbv7rrk+2xJA/7hw9g=
-modernc.org/cc/v4 v4.27.1 h1:9W30zRlYrefrDV2JE2O8VDtJ1yPGownxciz5rrbQZis=
+modernc.org/cc/v4 v4.26.5 h1:xM3bX7Mve6G8K8b+T11ReenJOT+BmVqQj0FY5T4+5Y4=
-modernc.org/cc/v4 v4.27.1/go.mod h1:uVtb5OGqUKpoLWhqwNQo/8LwvoiEBLvZXIQ/SmO6mL0=
+modernc.org/cc/v4 v4.26.5/go.mod h1:uVtb5OGqUKpoLWhqwNQo/8LwvoiEBLvZXIQ/SmO6mL0=
-modernc.org/ccgo/v4 v4.30.1 h1:4r4U1J6Fhj98NKfSjnPUN7Ze2c6MnAdL0hWw6+LrJpc=
+modernc.org/ccgo/v4 v4.28.1 h1:wPKYn5EC/mYTqBO373jKjvX2n+3+aK7+sICCv4Fjy1A=
-modernc.org/ccgo/v4 v4.30.1/go.mod h1:bIOeI1JL54Utlxn+LwrFyjCx2n2RDiYEaJVSrgdrRfM=
+modernc.org/ccgo/v4 v4.28.1/go.mod h1:uD+4RnfrVgE6ec9NGguUNdhqzNIeeomeXf6CL0GTE5Q=
 modernc.org/fileutil v1.3.40 h1:ZGMswMNc9JOCrcrakF1HrvmergNLAmxOPjizirpfqBA=
 modernc.org/fileutil v1.3.40/go.mod h1:HxmghZSZVAz/LXcMNwZPA/DRrQZEVP9VX0V4LQGQFOc=
 modernc.org/gc/v2 v2.6.5 h1:nyqdV8q46KvTpZlsw66kWqwXRHdjIlJOhG6kxiV/9xI=
 modernc.org/gc/v2 v2.6.5/go.mod h1:YgIahr1ypgfe7chRuJi2gD7DBQiKSLMPgBQe9oIiito=
-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 v1.66.10 h1:yZkb3YeLx4oynyR+iUsXsybsX4Ubx7MQlSYEw4yj59A=
 modernc.org/libc v1.66.10/go.mod h1:8vGSEwvoUoltr4dlywvHqjtAqHBaw0j1jI7iFBTAr2I=
-modernc.org/libc v1.67.1 h1:bFaqOaa5/zbWYJo8aW0tXPX21hXsngG2M7mckCnFSVk=
-modernc.org/libc v1.67.1/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=

internal/entities/system/system.go

@@ -155,16 +155,17 @@ type Info struct {
 
 // Data that does not change during process lifetime and is not needed in All Systems table
 type Details struct {
-	Hostname    string `cbor:"0,keyasint"`
+	Hostname      string        `cbor:"0,keyasint"`
-	Kernel      string `cbor:"1,keyasint,omitempty"`
+	Kernel        string        `cbor:"1,keyasint,omitempty"`
-	Cores       int    `cbor:"2,keyasint"`
+	Cores         int           `cbor:"2,keyasint"`
-	Threads     int    `cbor:"3,keyasint"`
+	Threads       int           `cbor:"3,keyasint"`
-	CpuModel    string `cbor:"4,keyasint"`
+	CpuModel      string        `cbor:"4,keyasint"`
-	Os          Os     `cbor:"5,keyasint"`
+	Os            Os            `cbor:"5,keyasint"`
-	OsName      string `cbor:"6,keyasint"`
+	OsName        string        `cbor:"6,keyasint"`
-	Arch        string `cbor:"7,keyasint"`
+	Arch          string        `cbor:"7,keyasint"`
-	Podman      bool   `cbor:"8,keyasint,omitempty"`
+	Podman        bool          `cbor:"8,keyasint,omitempty"`
-	MemoryTotal uint64 `cbor:"9,keyasint"`
+	MemoryTotal   uint64        `cbor:"9,keyasint"`
+	SmartInterval time.Duration `cbor:"10,keyasint,omitempty"`
 }
 
 // Final data structure to return to the hub

internal/hub/systems/system.go

@@ -42,8 +42,9 @@ type System struct {
 	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
-	smartFetched   atomic.Bool          // True if SMART devices 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
 }
 
 func (sm *SystemManager) NewSystem(systemId string) *System {
@@ -132,14 +133,19 @@ func (sys *System) update() error {
 	// create system records
 	_, err = sys.createRecords(data)
 
-	// Fetch and save SMART devices when system first comes online
+	// Fetch and save SMART devices when system first comes online or at intervals
-	if backgroundSmartFetchEnabled() && !sys.smartFetched.Load() && sys.smartFetching.CompareAndSwap(false, true) {
+	if backgroundSmartFetchEnabled() {
-		go func() {
+		if sys.smartInterval <= 0 {
-			defer sys.smartFetching.Store(false)
+			sys.smartInterval = time.Hour
-			if err := sys.FetchAndSaveSmartDevices(); err == nil {
+		}
-				sys.smartFetched.Store(true)
+		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
@@ -212,6 +218,10 @@ func (sys *System) createRecords(data *system.CombinedData) (*core.Record, error
 				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)

internal/site/src/components/routes/system/smart-table.tsx

@@ -93,51 +93,15 @@ export const smartColumns: ColumnDef<SmartAttribute>[] = [
 	},
 ]
 
-export type DiskInfo = {
-	id: string
-	system: string
-	device: string
-	model: string
-	capacity: string
-	status: string
-	temperature: number
-	deviceType: string
-	powerOnHours?: number
-	powerCycles?: number
-	attributes?: SmartAttribute[]
-	updated: string
-}
-
 // Function to format capacity display
 function formatCapacity(bytes: number): string {
 	const { value, unit } = formatBytes(bytes)
 	return `${toFixedFloat(value, value >= 10 ? 1 : 2)} ${unit}`
 }
 
-// Function to convert SmartDeviceRecord to DiskInfo
-function convertSmartDeviceRecordToDiskInfo(records: SmartDeviceRecord[]): DiskInfo[] {
-	const unknown = "Unknown"
-	return records.map((record) => ({
-		id: record.id,
-		system: record.system,
-		device: record.name || unknown,
-		model: record.model || unknown,
-		serialNumber: record.serial || unknown,
-		firmwareVersion: record.firmware || unknown,
-		capacity: record.capacity ? formatCapacity(record.capacity) : unknown,
-		status: record.state || unknown,
-		temperature: record.temp || 0,
-		deviceType: record.type || unknown,
-		attributes: record.attributes,
-		updated: record.updated,
-		powerOnHours: record.hours,
-		powerCycles: record.cycles,
-	}))
-}
-
 const SMART_DEVICE_FIELDS = "id,system,name,model,state,capacity,temp,type,hours,cycles,updated"
 
-export const columns: ColumnDef<DiskInfo>[] = [
+export const columns: ColumnDef<SmartDeviceRecord>[] = [
 	{
 		id: "system",
 		accessorFn: (record) => record.system,
@@ -154,12 +118,12 @@ export const columns: ColumnDef<DiskInfo>[] = [
 		},
 	},
 	{
-		accessorKey: "device",
+		accessorKey: "name",
-		sortingFn: (a, b) => a.original.device.localeCompare(b.original.device),
+		sortingFn: (a, b) => a.original.name.localeCompare(b.original.name),
 		header: ({ column }) => <HeaderButton column={column} name={t`Device`} Icon={HardDrive} />,
-		cell: ({ row }) => (
+		cell: ({ getValue }) => (
-			<div className="font-medium max-w-40 truncate ms-1.5" title={row.getValue("device")}>
+			<div className="font-medium max-w-40 truncate ms-1.5" title={getValue() as string}>
-				{row.getValue("device")}
+				{getValue() as string}
 			</div>
 		),
 	},
@@ -167,19 +131,20 @@ export const columns: ColumnDef<DiskInfo>[] = [
 		accessorKey: "model",
 		sortingFn: (a, b) => a.original.model.localeCompare(b.original.model),
 		header: ({ column }) => <HeaderButton column={column} name={t`Model`} Icon={Box} />,
-		cell: ({ row }) => (
+		cell: ({ getValue }) => (
-			<div className="max-w-48 truncate ms-1.5" title={row.getValue("model")}>
+			<div className="max-w-48 truncate ms-1.5" title={getValue() as string}>
-				{row.getValue("model")}
+				{getValue() as string}
 			</div>
 		),
 	},
 	{
 		accessorKey: "capacity",
+		invertSorting: true,
 		header: ({ column }) => <HeaderButton column={column} name={t`Capacity`} Icon={BinaryIcon} />,
-		cell: ({ getValue }) => <span className="ms-1.5">{getValue() as string}</span>,
+		cell: ({ getValue }) => <span className="ms-1.5">{formatCapacity(getValue() as number)}</span>,
 	},
 	{
-		accessorKey: "status",
+		accessorKey: "state",
 		header: ({ column }) => <HeaderButton column={column} name={t`Status`} Icon={Activity} />,
 		cell: ({ getValue }) => {
 			const status = getValue() as string
@@ -191,8 +156,8 @@ export const columns: ColumnDef<DiskInfo>[] = [
 		},
 	},
 	{
-		accessorKey: "deviceType",
+		accessorKey: "type",
-		sortingFn: (a, b) => a.original.deviceType.localeCompare(b.original.deviceType),
+		sortingFn: (a, b) => a.original.type.localeCompare(b.original.type),
 		header: ({ column }) => <HeaderButton column={column} name={t`Type`} Icon={ArrowLeftRightIcon} />,
 		cell: ({ getValue }) => (
 			<div className="ms-1.5">
@@ -203,7 +168,7 @@ export const columns: ColumnDef<DiskInfo>[] = [
 		),
 	},
 	{
-		accessorKey: "powerOnHours",
+		accessorKey: "hours",
 		invertSorting: true,
 		header: ({ column }) => (
 			<HeaderButton column={column} name={t({ message: "Power On", comment: "Power On Time" })} Icon={Clock} />
@@ -223,7 +188,7 @@ export const columns: ColumnDef<DiskInfo>[] = [
 		},
 	},
 	{
-		accessorKey: "powerCycles",
+		accessorKey: "cycles",
 		invertSorting: true,
 		header: ({ column }) => (
 			<HeaderButton column={column} name={t({ message: "Cycles", comment: "Power Cycles" })} Icon={RotateCwIcon} />
@@ -237,7 +202,7 @@ export const columns: ColumnDef<DiskInfo>[] = [
 		},
 	},
 	{
-		accessorKey: "temperature",
+		accessorKey: "temp",
 		invertSorting: true,
 		header: ({ column }) => <HeaderButton column={column} name={t`Temp`} Icon={ThermometerIcon} />,
 		cell: ({ getValue }) => {
@@ -246,14 +211,14 @@ export const columns: ColumnDef<DiskInfo>[] = [
 		},
 	},
 	// {
-	// 	accessorKey: "serialNumber",
+	// 	accessorKey: "serial",
-	// 	sortingFn: (a, b) => a.original.serialNumber.localeCompare(b.original.serialNumber),
+	// 	sortingFn: (a, b) => a.original.serial.localeCompare(b.original.serial),
 	// 	header: ({ column }) => <HeaderButton column={column} name={t`Serial Number`} Icon={HashIcon} />,
 	// 	cell: ({ getValue }) => <span className="ms-1.5">{getValue() as string}</span>,
 	// },
 	// {
-	// 	accessorKey: "firmwareVersion",
+	// 	accessorKey: "firmware",
-	// 	sortingFn: (a, b) => a.original.firmwareVersion.localeCompare(b.original.firmwareVersion),
+	// 	sortingFn: (a, b) => a.original.firmware.localeCompare(b.original.firmware),
 	// 	header: ({ column }) => <HeaderButton column={column} name={t`Firmware`} Icon={CpuIcon} />,
 	// 	cell: ({ getValue }) => <span className="ms-1.5">{getValue() as string}</span>,
 	// },
@@ -272,7 +237,15 @@ export const columns: ColumnDef<DiskInfo>[] = [
 	},
 ]
 
-function HeaderButton({ column, name, Icon }: { column: Column<DiskInfo>; name: string; Icon: React.ElementType }) {
+function HeaderButton({
+	column,
+	name,
+	Icon,
+}: {
+	column: Column<SmartDeviceRecord>
+	name: string
+	Icon: React.ElementType
+}) {
 	const isSorted = column.getIsSorted()
 	return (
 		<Button
@@ -290,7 +263,7 @@ function HeaderButton({ column, name, Icon }: { column: Column<DiskInfo>; name:
 }
 
 export default function DisksTable({ systemId }: { systemId?: string }) {
-	const [sorting, setSorting] = useState<SortingState>([{ id: systemId ? "device" : "system", desc: false }])
+	const [sorting, setSorting] = useState<SortingState>([{ id: systemId ? "name" : "system", desc: false }])
 	const [columnFilters, setColumnFilters] = useState<ColumnFiltersState>([])
 	const [rowSelection, setRowSelection] = useState({})
 	const [smartDevices, setSmartDevices] = useState<SmartDeviceRecord[] | undefined>(undefined)
@@ -299,96 +272,95 @@ export default function DisksTable({ systemId }: { systemId?: string }) {
 	const [rowActionState, setRowActionState] = useState<{ type: "refresh" | "delete"; id: string } | null>(null)
 	const [globalFilter, setGlobalFilter] = useState("")
 
-	const openSheet = (disk: DiskInfo) => {
+	const openSheet = (disk: SmartDeviceRecord) => {
 		setActiveDiskId(disk.id)
 		setSheetOpen(true)
 	}
 
-	// Fetch smart devices from collection (without attributes to save bandwidth)
+	// Fetch smart devices
-	const fetchSmartDevices = useCallback(() => {
+	useEffect(() => {
+		const controller = new AbortController()
+
 		pb.collection<SmartDeviceRecord>("smart_devices")
 			.getFullList({
 				filter: systemId ? pb.filter("system = {:system}", { system: systemId }) : undefined,
 				fields: SMART_DEVICE_FIELDS,
+				signal: controller.signal,
 			})
-			.then((records) => {
+			.then(setSmartDevices)
-				setSmartDevices(records)
+			.catch((err) => {
+				if (!err.isAbort) {
+					setSmartDevices([])
+				}
 			})
-			.catch(() => setSmartDevices([]))
+
+		return () => controller.abort()
 	}, [systemId])
 
-	// Fetch smart devices when component mounts or systemId changes
+	// Subscribe to updates
-	useEffect(() => {
-		fetchSmartDevices()
-	}, [fetchSmartDevices])
-
-	// Subscribe to live updates so rows add/remove without manual refresh/filtering
 	useEffect(() => {
 		let unsubscribe: (() => void) | undefined
 		const pbOptions = systemId
 			? { fields: SMART_DEVICE_FIELDS, filter: pb.filter("system = {:system}", { system: systemId }) }
 			: { fields: SMART_DEVICE_FIELDS }
 
-		;(async () => {
+			; (async () => {
-			try {
+				try {
-				unsubscribe = await pb.collection("smart_devices").subscribe(
+					unsubscribe = await pb.collection("smart_devices").subscribe(
-					"*",
+						"*",
-					(event) => {
+						(event) => {
-						const record = event.record as SmartDeviceRecord
+							const record = event.record as SmartDeviceRecord
-						setSmartDevices((currentDevices) => {
+							setSmartDevices((currentDevices) => {
-							const devices = currentDevices ?? []
+								const devices = currentDevices ?? []
-							const matchesSystemScope = !systemId || record.system === systemId
+								const matchesSystemScope = !systemId || record.system === systemId
 
-							if (event.action === "delete") {
+								if (event.action === "delete") {
-								return devices.filter((device) => device.id !== record.id)
+									return devices.filter((device) => device.id !== record.id)
-							}
+								}
 
-							if (!matchesSystemScope) {
+								if (!matchesSystemScope) {
-								// Record moved out of scope; ensure it disappears locally.
+									// Record moved out of scope; ensure it disappears locally.
-								return devices.filter((device) => device.id !== record.id)
+									return devices.filter((device) => device.id !== record.id)
-							}
+								}
 
-							const existingIndex = devices.findIndex((device) => device.id === record.id)
+								const existingIndex = devices.findIndex((device) => device.id === record.id)
-							if (existingIndex === -1) {
+								if (existingIndex === -1) {
-								return [record, ...devices]
+									return [record, ...devices]
-							}
+								}
 
-							const next = [...devices]
+								const next = [...devices]
-							next[existingIndex] = record
+								next[existingIndex] = record
-							return next
+								return next
-						})
+							})
-					},
+						},
-					pbOptions
+						pbOptions
-				)
+					)
-			} catch (error) {
+				} catch (error) {
-				console.error("Failed to subscribe to SMART device updates:", error)
+					console.error("Failed to subscribe to SMART device updates:", error)
-			}
+				}
-		})()
+			})()
 
 		return () => {
 			unsubscribe?.()
 		}
 	}, [systemId])
 
-	const handleRowRefresh = useCallback(
+	const handleRowRefresh = useCallback(async (disk: SmartDeviceRecord) => {
-		async (disk: DiskInfo) => {
+		if (!disk.system) return
-			if (!disk.system) return
+		setRowActionState({ type: "refresh", id: disk.id })
-			setRowActionState({ type: "refresh", id: disk.id })
+		try {
-			try {
+			await pb.send("/api/beszel/smart/refresh", {
-				await pb.send("/api/beszel/smart/refresh", {
+				method: "POST",
-					method: "POST",
+				query: { system: disk.system },
-					query: { system: disk.system },
+			})
-				})
+		} catch (error) {
-			} catch (error) {
+			console.error("Failed to refresh SMART device:", error)
-				console.error("Failed to refresh SMART device:", error)
+		} finally {
-			} finally {
+			setRowActionState((state) => (state?.id === disk.id ? null : state))
-				setRowActionState((state) => (state?.id === disk.id ? null : state))
+		}
-			}
+	}, [])
-		},
-		[fetchSmartDevices]
-	)
 
-	const handleDeleteDevice = useCallback(async (disk: DiskInfo) => {
+	const handleDeleteDevice = useCallback(async (disk: SmartDeviceRecord) => {
 		setRowActionState({ type: "delete", id: disk.id })
 		try {
 			await pb.collection("smart_devices").delete(disk.id)
@@ -400,7 +372,7 @@ export default function DisksTable({ systemId }: { systemId?: string }) {
 		}
 	}, [])
 
-	const actionColumn = useMemo<ColumnDef<DiskInfo>>(
+	const actionColumn = useMemo<ColumnDef<SmartDeviceRecord>>(
 		() => ({
 			id: "actions",
 			enableSorting: false,
@@ -468,13 +440,8 @@ export default function DisksTable({ systemId }: { systemId?: string }) {
 		return [...baseColumns, actionColumn]
 	}, [systemId, actionColumn])
 
-	// Convert SmartDeviceRecord to DiskInfo
-	const diskData = useMemo(() => {
-		return smartDevices ? convertSmartDeviceRecordToDiskInfo(smartDevices) : []
-	}, [smartDevices])
-
 	const table = useReactTable({
-		data: diskData,
+		data: smartDevices || ([] as SmartDeviceRecord[]),
 		columns: tableColumns,
 		onSortingChange: setSorting,
 		onColumnFiltersChange: setColumnFilters,
@@ -492,10 +459,10 @@ export default function DisksTable({ systemId }: { systemId?: string }) {
 		globalFilterFn: (row, _columnId, filterValue) => {
 			const disk = row.original
 			const systemName = $allSystemsById.get()[disk.system]?.name ?? ""
-			const device = disk.device ?? ""
+			const device = disk.name ?? ""
 			const model = disk.model ?? ""
-			const status = disk.status ?? ""
+			const status = disk.state ?? ""
-			const type = disk.deviceType ?? ""
+			const type = disk.type ?? ""
 			const searchString = `${systemName} ${device} ${model} ${status} ${type}`.toLowerCase()
 			return (filterValue as string)
 				.toLowerCase()
@@ -505,7 +472,7 @@ export default function DisksTable({ systemId }: { systemId?: string }) {
 	})
 
 	// Hide the table on system pages if there's no data, but always show on global page
-	if (systemId && !diskData.length && !columnFilters.length) {
+	if (systemId && !smartDevices?.length && !columnFilters.length) {
 		return null
 	}
 

supplemental/CHANGELOG.md

@@ -1,3 +1,15 @@
+## 0.18.0
+
+- Collect S.M.A.R.T. data in the background every hour.
+
+- Add `SMART_INTERVAL` environment variable to customize S.M.A.R.T. data collection interval.
+
+- Collect system distribution and architecture.
+
+- Add `system_details` collection to store infrequently updated system information.
+
+- Skip known non-unique product UUID when generating fingerprints. (#1556)
+
 ## 0.17.0
 
 - Add quiet hours to silence alerts during specific time periods. (#265)