[Feature] Add detailed CPU metrics (User, System, IOWait, Steal) with per-core monitoring (#1356)

* Add user, system io wait

* add per cpu core

* add total

agent/battery/battery.go

@@ -10,8 +10,10 @@ import (
 	"github.com/distatus/battery"
 )
 
-var systemHasBattery = false
-var haveCheckedBattery = false
+var (
+	systemHasBattery   = false
+	haveCheckedBattery = false
+)
 
 // HasReadableBattery checks if the system has a battery and returns true if it does.
 func HasReadableBattery() bool {
@@ -21,7 +23,7 @@ func HasReadableBattery() bool {
 	haveCheckedBattery = true
 	batteries, err := battery.GetAll()
 	for _, bat := range batteries {
-		if bat.Full > 0 {
+		if bat != nil && (bat.Full > 0 || bat.Design > 0) {
 			systemHasBattery = true
 			break
 		}

agent/cpu.go

@@ -8,6 +8,7 @@ import (
 )
 
 var lastCpuTimes = make(map[uint16]cpu.TimesStat)
+var lastPerCoreCpuTimes = make(map[uint16][]cpu.TimesStat)
 
 // init initializes the CPU monitoring by storing the initial CPU times
 // for the default 60-second cache interval.
@@ -15,6 +16,18 @@ func init() {
 	if times, err := cpu.Times(false); err == nil {
 		lastCpuTimes[60000] = times[0]
 	}
+	if perCoreTimes, err := cpu.Times(true); err == nil {
+		lastPerCoreCpuTimes[60000] = perCoreTimes
+	}
+}
+
+// CpuMetrics contains detailed CPU usage breakdown
+type CpuMetrics struct {
+	Total  float64
+	User   float64
+	System float64
+	Iowait float64
+	Steal  float64
 }
 
 // getCpuPercent calculates the CPU usage percentage using cached previous measurements.
@@ -34,6 +47,92 @@ func getCpuPercent(cacheTimeMs uint16) (float64, error) {
 	return delta, nil
 }
 
+// getCpuMetrics calculates detailed CPU usage metrics using cached previous measurements.
+// It returns percentages for total, user, system, iowait, and steal time.
+func getCpuMetrics(cacheTimeMs uint16) (CpuMetrics, error) {
+	times, err := cpu.Times(false)
+	if err != nil || len(times) == 0 {
+		return CpuMetrics{}, err
+	}
+	// if cacheTimeMs is not in lastCpuTimes, use 60000 as fallback lastCpuTime
+	if _, ok := lastCpuTimes[cacheTimeMs]; !ok {
+		lastCpuTimes[cacheTimeMs] = lastCpuTimes[60000]
+	}
+
+	t1 := lastCpuTimes[cacheTimeMs]
+	t2 := times[0]
+
+	t1All, t1Busy := getAllBusy(t1)
+	t2All, t2Busy := getAllBusy(t2)
+
+	totalDelta := t2All - t1All
+	if totalDelta <= 0 {
+		return CpuMetrics{}, nil
+	}
+
+	metrics := CpuMetrics{
+		Total:  clampPercent((t2Busy - t1Busy) / totalDelta * 100),
+		User:   clampPercent((t2.User - t1.User) / totalDelta * 100),
+		System: clampPercent((t2.System - t1.System) / totalDelta * 100),
+		Iowait: clampPercent((t2.Iowait - t1.Iowait) / totalDelta * 100),
+		Steal:  clampPercent((t2.Steal - t1.Steal) / totalDelta * 100),
+	}
+
+	lastCpuTimes[cacheTimeMs] = times[0]
+	return metrics, nil
+}
+
+// clampPercent ensures the percentage is between 0 and 100
+func clampPercent(value float64) float64 {
+	return math.Min(100, math.Max(0, value))
+}
+
+// getPerCoreCpuMetrics calculates per-core CPU usage metrics.
+// Returns a map where the key is "cpu0", "cpu1", etc. and the value is an array of [user, system, iowait, steal] percentages.
+func getPerCoreCpuMetrics(cacheTimeMs uint16) (map[string][4]float64, error) {
+	perCoreTimes, err := cpu.Times(true)
+	if err != nil || len(perCoreTimes) == 0 {
+		return nil, err
+	}
+
+	// Initialize cache if needed
+	if _, ok := lastPerCoreCpuTimes[cacheTimeMs]; !ok {
+		lastPerCoreCpuTimes[cacheTimeMs] = lastPerCoreCpuTimes[60000]
+	}
+
+	lastTimes := lastPerCoreCpuTimes[cacheTimeMs]
+	result := make(map[string][4]float64)
+
+	// Calculate metrics for each core
+	for i, currentTime := range perCoreTimes {
+		if i >= len(lastTimes) {
+			break
+		}
+
+		t1 := lastTimes[i]
+		t2 := currentTime
+
+		t1All, _ := getAllBusy(t1)
+		t2All, _ := getAllBusy(t2)
+
+		totalDelta := t2All - t1All
+		if totalDelta <= 0 {
+			continue
+		}
+
+		// Store as [user, system, iowait, steal]
+		result[currentTime.CPU] = [4]float64{
+			clampPercent((t2.User - t1.User) / totalDelta * 100),
+			clampPercent((t2.System - t1.System) / totalDelta * 100),
+			clampPercent((t2.Iowait - t1.Iowait) / totalDelta * 100),
+			clampPercent((t2.Steal - t1.Steal) / totalDelta * 100),
+		}
+	}
+
+	lastPerCoreCpuTimes[cacheTimeMs] = perCoreTimes
+	return result, nil
+}
+
 // calculateBusy calculates the CPU busy percentage between two time points.
 // It computes the ratio of busy time to total time elapsed between t1 and t2,
 // returning a percentage clamped between 0 and 100.

agent/docker.go

@@ -13,6 +13,7 @@ import (
 	"net/http"
 	"net/url"
 	"os"
+	"path"
 	"strings"
 	"sync"
 	"time"
@@ -53,6 +54,7 @@ type dockerManager struct {
 	buf                 *bytes.Buffer               // Buffer to store and read response bodies
 	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
 
 	// Cache-time-aware tracking for CPU stats (similar to cpu.go)
 	// Maps cache time intervals to container-specific CPU usage tracking
@@ -94,6 +96,19 @@ func (d *dockerManager) dequeue() {
 	}
 }
 
+// shouldExcludeContainer checks if a container name matches any exclusion pattern
+func (dm *dockerManager) shouldExcludeContainer(name string) bool {
+	if len(dm.excludeContainers) == 0 {
+		return false
+	}
+	for _, pattern := range dm.excludeContainers {
+		if match, _ := path.Match(pattern, name); match {
+			return true
+		}
+	}
+	return false
+}
+
 // Returns stats for all running containers with cache-time-aware delta tracking
 func (dm *dockerManager) getDockerStats(cacheTimeMs uint16) ([]*container.Stats, error) {
 	resp, err := dm.client.Get("http://localhost/containers/json")
@@ -121,6 +136,13 @@ func (dm *dockerManager) getDockerStats(cacheTimeMs uint16) ([]*container.Stats,
 
 	for _, ctr := range dm.apiContainerList {
 		ctr.IdShort = ctr.Id[:12]
+
+		// Skip this container if it matches the exclusion pattern
+		if dm.shouldExcludeContainer(ctr.Names[0][1:]) {
+			slog.Debug("Excluding container", "name", ctr.Names[0][1:])
+			continue
+		}
+
 		dm.validIds[ctr.IdShort] = struct{}{}
 		// check if container is less than 1 minute old (possible restart)
 		// note: can't use Created field because it's not updated on restart
@@ -503,6 +525,19 @@ func newDockerManager(a *Agent) *dockerManager {
 		userAgent: "Docker-Client/",
 	}
 
+	// Read container exclusion patterns from environment variable
+	var excludeContainers []string
+	if excludeStr, set := GetEnv("EXCLUDE_CONTAINERS"); set && excludeStr != "" {
+		parts := strings.SplitSeq(excludeStr, ",")
+		for part := range parts {
+			trimmed := strings.TrimSpace(part)
+			if trimmed != "" {
+				excludeContainers = append(excludeContainers, trimmed)
+			}
+		}
+		slog.Info("EXCLUDE_CONTAINERS", "patterns", excludeContainers)
+	}
+
 	manager := &dockerManager{
 		client: &http.Client{
 			Timeout:   timeout,
@@ -512,6 +547,7 @@ func newDockerManager(a *Agent) *dockerManager {
 		sem:               make(chan struct{}, 5),
 		apiContainerList:  []*container.ApiInfo{},
 		apiStats:          &container.ApiStats{},
+		excludeContainers: excludeContainers,
 
 		// Initialize cache-time-aware tracking structures
 		lastCpuContainer:    make(map[uint16]map[string]uint64),

agent/docker_test.go

@@ -1099,3 +1099,107 @@ func TestAllocateBuffer(t *testing.T) {
 		})
 	}
 }
+
+func TestShouldExcludeContainer(t *testing.T) {
+	tests := []struct {
+		name          string
+		containerName string
+		patterns      []string
+		expected      bool
+	}{
+		{
+			name:          "empty patterns excludes nothing",
+			containerName: "any-container",
+			patterns:      []string{},
+			expected:      false,
+		},
+		{
+			name:          "exact match - excluded",
+			containerName: "test-web",
+			patterns:      []string{"test-web", "test-api"},
+			expected:      true,
+		},
+		{
+			name:          "exact match - not excluded",
+			containerName: "prod-web",
+			patterns:      []string{"test-web", "test-api"},
+			expected:      false,
+		},
+		{
+			name:          "wildcard prefix match - excluded",
+			containerName: "test-web",
+			patterns:      []string{"test-*"},
+			expected:      true,
+		},
+		{
+			name:          "wildcard prefix match - not excluded",
+			containerName: "prod-web",
+			patterns:      []string{"test-*"},
+			expected:      false,
+		},
+		{
+			name:          "wildcard suffix match - excluded",
+			containerName: "myapp-staging",
+			patterns:      []string{"*-staging"},
+			expected:      true,
+		},
+		{
+			name:          "wildcard suffix match - not excluded",
+			containerName: "myapp-prod",
+			patterns:      []string{"*-staging"},
+			expected:      false,
+		},
+		{
+			name:          "wildcard both sides match - excluded",
+			containerName: "test-myapp-staging",
+			patterns:      []string{"*-myapp-*"},
+			expected:      true,
+		},
+		{
+			name:          "wildcard both sides match - not excluded",
+			containerName: "prod-yourapp-live",
+			patterns:      []string{"*-myapp-*"},
+			expected:      false,
+		},
+		{
+			name:          "multiple patterns - matches first",
+			containerName: "test-container",
+			patterns:      []string{"test-*", "*-staging"},
+			expected:      true,
+		},
+		{
+			name:          "multiple patterns - matches second",
+			containerName: "myapp-staging",
+			patterns:      []string{"test-*", "*-staging"},
+			expected:      true,
+		},
+		{
+			name:          "multiple patterns - no match",
+			containerName: "prod-web",
+			patterns:      []string{"test-*", "*-staging"},
+			expected:      false,
+		},
+		{
+			name:          "mixed exact and wildcard - exact match",
+			containerName: "temp-container",
+			patterns:      []string{"temp-container", "test-*"},
+			expected:      true,
+		},
+		{
+			name:          "mixed exact and wildcard - wildcard match",
+			containerName: "test-web",
+			patterns:      []string{"temp-container", "test-*"},
+			expected:      true,
+		},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			dm := &dockerManager{
+				excludeContainers: tt.patterns,
+			}
+			result := dm.shouldExcludeContainer(tt.containerName)
+			assert.Equal(t, tt.expected, result)
+		})
+	}
+}

agent/smart.go

@@ -426,7 +426,7 @@ func (sm *SmartManager) smartctlArgs(deviceInfo *DeviceInfo, includeStandby bool
 		}
 	}
 
-	args = append(args, "-aj")
+	args = append(args, "-a", "--json=c")
 
 	if includeStandby {
 		args = append(args, "-n", "standby")
@@ -688,13 +688,17 @@ func (sm *SmartManager) parseSmartForSata(output []byte) (bool, int) {
 	// update SmartAttributes
 	smartData.Attributes = make([]*smart.SmartAttribute, 0, len(data.AtaSmartAttributes.Table))
 	for _, attr := range data.AtaSmartAttributes.Table {
+		rawValue := uint64(attr.Raw.Value)
+		if parsed, ok := smart.ParseSmartRawValueString(attr.Raw.String); ok {
+			rawValue = parsed
+		}
 		smartAttr := &smart.SmartAttribute{
 			ID:         attr.ID,
 			Name:       attr.Name,
 			Value:      attr.Value,
 			Worst:      attr.Worst,
 			Threshold:  attr.Thresh,
-			RawValue:   uint64(attr.Raw.Value),
+			RawValue:   rawValue,
 			RawString:  attr.Raw.String,
 			WhenFailed: attr.WhenFailed,
 		}

agent/smart_test.go

@@ -89,6 +89,49 @@ func TestParseSmartForSata(t *testing.T) {
 	}
 }
 
+func TestParseSmartForSataParentheticalRawValue(t *testing.T) {
+	jsonPayload := []byte(`{
+		"smartctl": {"exit_status": 0},
+		"device": {"name": "/dev/sdz", "type": "sat"},
+		"model_name": "Example",
+		"serial_number": "PARENTHESES123",
+		"firmware_version": "1.0",
+		"user_capacity": {"bytes": 1024},
+		"smart_status": {"passed": true},
+		"temperature": {"current": 25},
+		"ata_smart_attributes": {
+			"table": [
+				{
+					"id": 9,
+					"name": "Power_On_Hours",
+					"value": 93,
+					"worst": 55,
+					"thresh": 0,
+					"when_failed": "",
+                    "raw": {
+                        "value": 57891864217128,
+                        "string": "39925 (212 206 0)"
+                    }
+				}
+			]
+		}
+	}`)
+
+	sm := &SmartManager{SmartDataMap: make(map[string]*smart.SmartData)}
+
+	hasData, exitStatus := sm.parseSmartForSata(jsonPayload)
+	require.True(t, hasData)
+	assert.Equal(t, 0, exitStatus)
+
+	data, ok := sm.SmartDataMap["PARENTHESES123"]
+	require.True(t, ok)
+	require.Len(t, data.Attributes, 1)
+
+	attr := data.Attributes[0]
+	assert.Equal(t, uint64(39925), attr.RawValue)
+	assert.Equal(t, "39925 (212 206 0)", attr.RawString)
+}
+
 func TestParseSmartForNvme(t *testing.T) {
 	fixturePath := filepath.Join("test-data", "smart", "nvme0.json")
 	data, err := os.ReadFile(fixturePath)
@@ -198,7 +241,7 @@ func TestSmartctlArgsWithoutType(t *testing.T) {
 	sm := &SmartManager{}
 
 	args := sm.smartctlArgs(device, true)
-	assert.Equal(t, []string{"-aj", "-n", "standby", "/dev/sda"}, args)
+	assert.Equal(t, []string{"-a", "--json=c", "-n", "standby", "/dev/sda"}, args)
 }
 
 func TestSmartctlArgs(t *testing.T) {
@@ -206,17 +249,17 @@ func TestSmartctlArgs(t *testing.T) {
 
 	sataDevice := &DeviceInfo{Name: "/dev/sda", Type: "sat"}
 	assert.Equal(t,
-		[]string{"-d", "sat", "-aj", "-n", "standby", "/dev/sda"},
+		[]string{"-d", "sat", "-a", "--json=c", "-n", "standby", "/dev/sda"},
 		sm.smartctlArgs(sataDevice, true),
 	)
 
 	assert.Equal(t,
-		[]string{"-d", "sat", "-aj", "/dev/sda"},
+		[]string{"-d", "sat", "-a", "--json=c", "/dev/sda"},
 		sm.smartctlArgs(sataDevice, false),
 	)
 
 	assert.Equal(t,
-		[]string{"-aj", "-n", "standby"},
+		[]string{"-a", "--json=c", "-n", "standby"},
 		sm.smartctlArgs(nil, true),
 	)
 }

agent/system.go

@@ -83,12 +83,21 @@ func (a *Agent) getSystemStats(cacheTimeMs uint16) system.Stats {
 		systemStats.Battery[1] = batteryState
 	}
 
-	// cpu percent
-	cpuPercent, err := getCpuPercent(cacheTimeMs)
+	// cpu metrics
+	cpuMetrics, err := getCpuMetrics(cacheTimeMs)
 	if err == nil {
-		systemStats.Cpu = twoDecimals(cpuPercent)
+		systemStats.Cpu = twoDecimals(cpuMetrics.Total)
+		systemStats.CpuUser = twoDecimals(cpuMetrics.User)
+		systemStats.CpuSystem = twoDecimals(cpuMetrics.System)
+		systemStats.CpuIowait = twoDecimals(cpuMetrics.Iowait)
+		systemStats.CpuSteal = twoDecimals(cpuMetrics.Steal)
 	} else {
-		slog.Error("Error getting cpu percent", "err", err)
+		slog.Error("Error getting cpu metrics", "err", err)
+	}
+
+	// per-core cpu metrics
+	if perCoreCpuMetrics, err := getPerCoreCpuMetrics(cacheTimeMs); err == nil && len(perCoreCpuMetrics) > 0 {
+		systemStats.CpuCores = perCoreCpuMetrics
 	}
 
 	// load average

internal/entities/smart/smart.go

@@ -1,6 +1,7 @@
 package smart
 
 import (
+	"encoding/json"
 	"strconv"
 	"strings"
 )
@@ -160,6 +161,33 @@ type RawValue struct {
 	String string        `json:"string"`
 }
 
+func (r *RawValue) UnmarshalJSON(data []byte) error {
+	var tmp struct {
+		Value  json.RawMessage `json:"value"`
+		String string          `json:"string"`
+	}
+
+	if err := json.Unmarshal(data, &tmp); err != nil {
+		return err
+	}
+
+	if len(tmp.Value) > 0 {
+		if err := r.Value.UnmarshalJSON(tmp.Value); err != nil {
+			return err
+		}
+	} else {
+		r.Value = 0
+	}
+
+	r.String = tmp.String
+
+	if parsed, ok := ParseSmartRawValueString(tmp.String); ok {
+		r.Value = SmartRawValue(parsed)
+	}
+
+	return nil
+}
+
 type SmartRawValue uint64
 
 // handles when drives report strings like "0h+0m+0.000s" or "7344 (253d 8h)" for power on hours
@@ -170,61 +198,73 @@ func (v *SmartRawValue) UnmarshalJSON(data []byte) error {
 		return nil
 	}
 
-	if trimmed[0] != '"' {
-		parsed, err := strconv.ParseUint(trimmed, 0, 64)
+	if trimmed[0] == '"' {
+		valueStr, err := strconv.Unquote(trimmed)
 		if err != nil {
 			return err
 		}
-		*v = SmartRawValue(parsed)
-		return nil
-	}
-
-	valueStr, err := strconv.Unquote(trimmed)
-	if err != nil {
-		return err
-	}
-	if valueStr == "" {
+		parsed, ok := ParseSmartRawValueString(valueStr)
+		if ok {
+			*v = SmartRawValue(parsed)
+			return nil
+		}
 		*v = 0
 		return nil
 	}
 
-	if parsed, err := strconv.ParseUint(valueStr, 0, 64); err == nil {
+	if parsed, err := strconv.ParseUint(trimmed, 0, 64); err == nil {
 		*v = SmartRawValue(parsed)
 		return nil
 	}
 
-	if idx := strings.IndexRune(valueStr, 'h'); idx >= 0 {
-		hoursPart := strings.TrimSpace(valueStr[:idx])
-		if hoursPart == "" {
-			*v = 0
-			return nil
-		}
-		if parsed, err := strconv.ParseFloat(hoursPart, 64); err == nil {
-			*v = SmartRawValue(uint64(parsed))
-			return nil
-		}
-	}
-
-	if digits := leadingDigitPrefix(valueStr); digits != "" {
-		if parsed, err := strconv.ParseUint(digits, 0, 64); err == nil {
-			*v = SmartRawValue(parsed)
-			return nil
-		}
+	if parsed, ok := ParseSmartRawValueString(trimmed); ok {
+		*v = SmartRawValue(parsed)
+		return nil
 	}
 
 	*v = 0
 	return nil
 }
 
-func leadingDigitPrefix(value string) string {
-	var builder strings.Builder
-	for _, r := range value {
-		if r < '0' || r > '9' {
-			break
-		}
-		builder.WriteRune(r)
+// ParseSmartRawValueString attempts to extract a numeric value from the raw value
+// strings emitted by smartctl, which sometimes include human-friendly annotations
+// like "7344 (253d 8h)" or "0h+0m+0.000s". It returns the parsed value and a
+// boolean indicating success.
+func ParseSmartRawValueString(value string) (uint64, bool) {
+	value = strings.TrimSpace(value)
+	if value == "" {
+		return 0, false
 	}
-	return builder.String()
+
+	if parsed, err := strconv.ParseUint(value, 0, 64); err == nil {
+		return parsed, true
+	}
+
+	if idx := strings.IndexRune(value, 'h'); idx > 0 {
+		hoursPart := strings.TrimSpace(value[:idx])
+		if hoursPart != "" {
+			if parsed, err := strconv.ParseFloat(hoursPart, 64); err == nil {
+				return uint64(parsed), true
+			}
+		}
+	}
+
+	for i := 0; i < len(value); i++ {
+		if value[i] < '0' || value[i] > '9' {
+			continue
+		}
+		end := i + 1
+		for end < len(value) && value[end] >= '0' && value[end] <= '9' {
+			end++
+		}
+		digits := value[i:end]
+		if parsed, err := strconv.ParseUint(digits, 10, 64); err == nil {
+			return parsed, true
+		}
+		i = end
+	}
+
+	return 0, false
 }
 
 // type PowerOnTimeInfo struct {

internal/entities/smart/smart_test.go

@@ -3,28 +3,60 @@ package smart
 import (
 	"encoding/json"
 	"testing"
+
+	"github.com/stretchr/testify/assert"
 )
 
 func TestSmartRawValueUnmarshalDuration(t *testing.T) {
 	input := []byte(`{"value":"62312h+33m+50.907s","string":"62312h+33m+50.907s"}`)
 	var raw RawValue
-	if err := json.Unmarshal(input, &raw); err != nil {
-		t.Fatalf("unexpected error unmarshalling raw value: %v", err)
-	}
+	err := json.Unmarshal(input, &raw)
+	assert.NoError(t, err)
 
-	if uint64(raw.Value) != 62312 {
-		t.Fatalf("expected hours to be 62312, got %d", raw.Value)
-	}
+	assert.EqualValues(t, 62312, raw.Value)
 }
 
 func TestSmartRawValueUnmarshalNumericString(t *testing.T) {
 	input := []byte(`{"value":"7344","string":"7344"}`)
 	var raw RawValue
-	if err := json.Unmarshal(input, &raw); err != nil {
-		t.Fatalf("unexpected error unmarshalling numeric string: %v", err)
-	}
+	err := json.Unmarshal(input, &raw)
+	assert.NoError(t, err)
 
-	if uint64(raw.Value) != 7344 {
-		t.Fatalf("expected hours to be 7344, got %d", raw.Value)
-	}
+	assert.EqualValues(t, 7344, raw.Value)
+}
+
+func TestSmartRawValueUnmarshalParenthetical(t *testing.T) {
+	input := []byte(`{"value":"39925 (212 206 0)","string":"39925 (212 206 0)"}`)
+	var raw RawValue
+	err := json.Unmarshal(input, &raw)
+	assert.NoError(t, err)
+
+	assert.EqualValues(t, 39925, raw.Value)
+}
+
+func TestSmartRawValueUnmarshalDurationWithFractions(t *testing.T) {
+	input := []byte(`{"value":"2748h+31m+49.560s","string":"2748h+31m+49.560s"}`)
+	var raw RawValue
+	err := json.Unmarshal(input, &raw)
+	assert.NoError(t, err)
+
+	assert.EqualValues(t, 2748, raw.Value)
+}
+
+func TestSmartRawValueUnmarshalParentheticalRawValue(t *testing.T) {
+	input := []byte(`{"value":57891864217128,"string":"39925 (212 206 0)"}`)
+	var raw RawValue
+	err := json.Unmarshal(input, &raw)
+	assert.NoError(t, err)
+
+	assert.EqualValues(t, 39925, raw.Value)
+}
+
+func TestSmartRawValueUnmarshalDurationRawValue(t *testing.T) {
+	input := []byte(`{"value":57891864217128,"string":"2748h+31m+49.560s"}`)
+	var raw RawValue
+	err := json.Unmarshal(input, &raw)
+	assert.NoError(t, err)
+
+	assert.EqualValues(t, 2748, raw.Value)
 }

internal/entities/system/system.go

@@ -11,7 +11,12 @@ import (
 type Stats struct {
 	Cpu            float64             `json:"cpu" cbor:"0,keyasint"`
 	MaxCpu         float64             `json:"cpum,omitempty" cbor:"1,keyasint,omitempty"`
-	Mem            float64             `json:"m" cbor:"2,keyasint"`
+	CpuUser        float64                `json:"cpuu,omitempty" cbor:"33,keyasint,omitempty"`
+	CpuSystem      float64                `json:"cpus,omitempty" cbor:"34,keyasint,omitempty"`
+	CpuIowait      float64                `json:"cpui,omitempty" cbor:"35,keyasint,omitempty"`
+	CpuSteal       float64                `json:"cpust,omitempty" cbor:"36,keyasint,omitempty"`
+	CpuCores       map[string][4]float64  `json:"cpuc,omitempty" cbor:"37,keyasint,omitempty"` // [user, system, iowait, steal] per core
+	Mem            float64                `json:"m" cbor:"2,keyasint"`
 	MemUsed        float64             `json:"mu" cbor:"3,keyasint"`
 	MemPct         float64             `json:"mp" cbor:"4,keyasint"`
 	MemBuffCache   float64             `json:"mb" cbor:"5,keyasint"`

internal/records/records.go

@@ -269,6 +269,10 @@ func (rm *RecordManager) AverageSystemStats(db dbx.Builder, records RecordIds) *
 				fs.DiskReadPs += value.DiskReadPs
 				fs.MaxDiskReadPS = max(fs.MaxDiskReadPS, value.MaxDiskReadPS, value.DiskReadPs)
 				fs.MaxDiskWritePS = max(fs.MaxDiskWritePS, value.MaxDiskWritePS, value.DiskWritePs)
+				fs.DiskReadBytes += value.DiskReadBytes
+				fs.DiskWriteBytes += value.DiskWriteBytes
+				fs.MaxDiskReadBytes = max(fs.MaxDiskReadBytes, value.MaxDiskReadBytes, value.DiskReadBytes)
+				fs.MaxDiskWriteBytes = max(fs.MaxDiskWriteBytes, value.MaxDiskWriteBytes, value.DiskWriteBytes)
 			}
 		}
 
@@ -356,6 +360,8 @@ func (rm *RecordManager) AverageSystemStats(db dbx.Builder, records RecordIds) *
 				fs.DiskUsed = twoDecimals(fs.DiskUsed / count)
 				fs.DiskWritePs = twoDecimals(fs.DiskWritePs / count)
 				fs.DiskReadPs = twoDecimals(fs.DiskReadPs / count)
+				fs.DiskReadBytes = fs.DiskReadBytes / uint64(count)
+				fs.DiskWriteBytes = fs.DiskWriteBytes / uint64(count)
 			}
 		}
 

internal/site/lingui.config.ts

@@ -11,6 +11,7 @@ export default defineConfig({
 		"es",
 		"fa",
 		"fr",
+		"he",
 		"hr",
 		"hu",
 		"it",

internal/site/src/components/charts/hooks.ts

@@ -118,6 +118,28 @@ export function useNetworkInterfaces(interfaces: SystemStats["ni"]) {
 				dataKey: ({ stats }: SystemStatsRecord) => stats?.ni?.[key]?.[index],
 				color: `hsl(${220 + (((sortedKeys.indexOf(key) * 360) / sortedKeys.length) % 360)}, 70%, 50%)`,
 
+				opacity: 0.3,
+			}))
+		},
+	}
+}
+
+// Assures consistent colors for CPU cores
+export function useCpuCores(cores: SystemStats["cpuc"]) {
+	const keys = Object.keys(cores ?? {})
+	// Sort cores by name (cpu0, cpu1, cpu2, etc.)
+	const sortedKeys = keys.sort((a, b) => {
+		const numA = Number.parseInt(a.replace("cpu", ""))
+		const numB = Number.parseInt(b.replace("cpu", ""))
+		return numA - numB
+	})
+	return {
+		length: sortedKeys.length,
+		data: (index = 0) => {
+			return sortedKeys.map((key) => ({
+				label: key,
+				dataKey: ({ stats }: SystemStatsRecord) => stats?.cpuc?.[key]?.[index],
+				color: `hsl(${(((sortedKeys.indexOf(key) * 360) / sortedKeys.length) % 360)}, 70%, 50%)`,
 				opacity: 0.3,
 			}))
 		},

internal/site/src/components/routes/system.tsx

@@ -73,6 +73,7 @@ import { Select, SelectContent, SelectItem, SelectTrigger, SelectValue } from ".
 import { Separator } from "../ui/separator"
 import { Tooltip, TooltipContent, TooltipProvider, TooltipTrigger } from "../ui/tooltip"
 import NetworkSheet from "./system/network-sheet"
+import CpuCoresSheet from "./system/cpu-cores-sheet"
 import LineChartDefault from "../charts/line-chart"
 
 
@@ -585,18 +586,49 @@ export default memo(function SystemDetail({ id }: { id: string }) {
 						grid={grid}
 						title={t`CPU Usage`}
 						description={t`Average system-wide CPU utilization`}
-						cornerEl={maxValSelect}
+						cornerEl={
+							<>
+								{maxValSelect}
+								<CpuCoresSheet chartData={chartData} dataEmpty={dataEmpty} grid={grid} maxValues={maxValues} />
+							</>
+						}
+						legend={true}
 					>
 						<AreaChartDefault
 							chartData={chartData}
 							maxToggled={maxValues}
+							legend={true}
 							dataPoints={[
 								{
-									label: t`CPU Usage`,
+									label: t`Total`,
 									dataKey: ({ stats }) => (showMax ? stats?.cpum : stats?.cpu),
 									color: 1,
 									opacity: 0.4,
 								},
+								{
+									label: t`User`,
+									dataKey: ({ stats }) => stats?.cpuu,
+									color: 2,
+									opacity: 0.3,
+								},
+								{
+									label: t`System`,
+									dataKey: ({ stats }) => stats?.cpus,
+									color: 3,
+									opacity: 0.3,
+								},
+								{
+									label: t`IOWait`,
+									dataKey: ({ stats }) => stats?.cpui,
+									color: 4,
+									opacity: 0.3,
+								},
+								{
+									label: t`Steal`,
+									dataKey: ({ stats }) => stats?.cpust,
+									color: 5,
+									opacity: 0.3,
+								},
 							]}
 							tickFormatter={(val) => `${toFixedFloat(val, 2)}%`}
 							contentFormatter={({ value }) => `${decimalString(value)}%`}
@@ -958,9 +990,9 @@ export default memo(function SystemDetail({ id }: { id: string }) {
 													label: t`Write`,
 													dataKey: ({ stats }) => {
 														if (showMax) {
-															return stats?.efs?.[extraFsName]?.wb ?? (stats?.efs?.[extraFsName]?.wm ?? 0) * 1024 * 1024
+															return stats?.efs?.[extraFsName]?.wbm || (stats?.efs?.[extraFsName]?.wm ?? 0) * 1024 * 1024
 														}
-														return stats?.efs?.[extraFsName]?.wb ?? (stats?.efs?.[extraFsName]?.w ?? 0) * 1024 * 1024
+														return stats?.efs?.[extraFsName]?.wb || (stats?.efs?.[extraFsName]?.w ?? 0) * 1024 * 1024
 													},
 													color: 3,
 													opacity: 0.3,

internal/site/src/components/routes/system/cpu-cores-sheet.tsx

@@ -0,0 +1,119 @@
+import { t } from "@lingui/core/macro"
+import { MoreHorizontalIcon } from "lucide-react"
+import { memo, useRef, useState } from "react"
+import AreaChartDefault from "@/components/charts/area-chart"
+import ChartTimeSelect from "@/components/charts/chart-time-select"
+import { Button } from "@/components/ui/button"
+import { Sheet, SheetContent, SheetTrigger } from "@/components/ui/sheet"
+import { DialogTitle } from "@/components/ui/dialog"
+import { decimalString, toFixedFloat } from "@/lib/utils"
+import type { ChartData, SystemStatsRecord } from "@/types"
+import { ChartCard } from "../system"
+
+export default memo(function CpuCoresSheet({
+	chartData,
+	dataEmpty,
+	grid,
+	maxValues,
+}: {
+	chartData: ChartData
+	dataEmpty: boolean
+	grid: boolean
+	maxValues: boolean
+}) {
+	const [cpuCoresOpen, setCpuCoresOpen] = useState(false)
+	const hasOpened = useRef(false)
+
+	if (cpuCoresOpen && !hasOpened.current) {
+		hasOpened.current = true
+	}
+
+	// Get list of CPU cores from the latest stats
+	const cpuCoresData = chartData.systemStats.at(-1)?.stats?.cpuc ?? {}
+	const coreNames = Object.keys(cpuCoresData).sort((a, b) => {
+		const numA = Number.parseInt(a.replace("cpu", ""))
+		const numB = Number.parseInt(b.replace("cpu", ""))
+		return numA - numB
+	})
+
+	if (coreNames.length === 0) {
+		return null
+	}
+
+	return (
+		<Sheet open={cpuCoresOpen} onOpenChange={setCpuCoresOpen}>
+			<DialogTitle className="sr-only">{t`Per-core CPU usage`}</DialogTitle>
+			<SheetTrigger asChild>
+				<Button
+					title={t`View per-core CPU`}
+					variant="outline"
+					size="icon"
+					className="shrink-0 max-sm:absolute max-sm:top-3 max-sm:end-3"
+				>
+					<MoreHorizontalIcon />
+				</Button>
+			</SheetTrigger>
+			{hasOpened.current && (
+				<SheetContent aria-describedby={undefined} className="overflow-auto w-200 !max-w-full p-4 sm:p-6">
+					<ChartTimeSelect className="w-[calc(100%-2em)]" agentVersion={chartData.agentVersion} />
+					{coreNames.map((coreName) => (
+						<ChartCard
+							key={coreName}
+							empty={dataEmpty}
+							grid={grid}
+							title={coreName.toUpperCase()}
+							description={t`CPU usage breakdown for ${coreName}`}
+							legend={true}
+							className="min-h-auto"
+						>
+							<AreaChartDefault
+								chartData={chartData}
+								maxToggled={maxValues}
+								legend={true}
+								dataPoints={[
+									{
+										label: t`Total`,
+										dataKey: ({ stats }: SystemStatsRecord) => {
+											const core = stats?.cpuc?.[coreName]
+											if (!core) return undefined
+											// Sum all metrics: user + system + iowait + steal
+											return core[0] + core[1] + core[2] + core[3]
+										},
+										color: 1,
+										opacity: 0.4,
+									},
+									{
+										label: t`User`,
+										dataKey: ({ stats }: SystemStatsRecord) => stats?.cpuc?.[coreName]?.[0],
+										color: 2,
+										opacity: 0.3,
+									},
+									{
+										label: t`System`,
+										dataKey: ({ stats }: SystemStatsRecord) => stats?.cpuc?.[coreName]?.[1],
+										color: 3,
+										opacity: 0.3,
+									},
+									{
+										label: t`IOWait`,
+										dataKey: ({ stats }: SystemStatsRecord) => stats?.cpuc?.[coreName]?.[2],
+										color: 4,
+										opacity: 0.3,
+									},
+									{
+										label: t`Steal`,
+										dataKey: ({ stats }: SystemStatsRecord) => stats?.cpuc?.[coreName]?.[3],
+										color: 5,
+										opacity: 0.3,
+									},
+								]}
+								tickFormatter={(val) => `${toFixedFloat(val, 2)}%`}
+								contentFormatter={({ value }) => `${decimalString(value)}%`}
+							/>
+						</ChartCard>
+					))}
+				</SheetContent>
+			)}
+		</Sheet>
+	)
+})

internal/site/src/types.d.ts

@@ -84,6 +84,16 @@ export interface SystemStats {
 	cpu: number
 	/** peak cpu */
 	cpum?: number
+	/** cpu user percent */
+	cpuu?: number
+	/** cpu system percent */
+	cpus?: number
+	/** cpu iowait percent */
+	cpui?: number
+	/** cpu steal percent */
+	cpust?: number
+	/** per-core cpu metrics [user, system, iowait, steal] */
+	cpuc?: Record<string, [number, number, number, number]>
 	// TODO: remove these in future release in favor of la
 	/** load average 1 minute */
 	l1?: number

supplemental/CHANGELOG.md

@@ -1,3 +1,11 @@
+## 0.15.3
+
+- Improve parsing of edge case S.M.A.R.T. power on times. (#1347)
+
+## 0.15.2
+
+- Improve S.M.A.R.T. device detection logic (fix regression in 0.15.1) (#1345)
+
 ## 0.15.1
 
 - Add `SMART_DEVICES` environment variable to specify devices and types. (#373, #1335)