add experimental nvml gpu collector (#1522)

- also update bun.lock

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) {
@@ -103,7 +104,17 @@ func NewAgent(dataDir ...string) (agent *Agent, err error) {
 	agent.dockerManager = newDockerManager()
 
 	// initialize system info
-	agent.refreshStaticInfo()
+	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)
@@ -166,7 +177,7 @@ func (a *Agent) gatherStats(options common.DataRequestOptions) *system.CombinedD
 		Info:  a.systemInfo,
 	}
 
-	// Include static info only when requested
+	// Include static system details only when requested
 	if options.IncludeDetails {
 		data.Details = &a.systemDetails
 	}
@@ -233,7 +244,8 @@ func (a *Agent) getFingerprint() string {
 
 	// if no fingerprint is found, generate one
 	fingerprint, err := host.HostID()
-	if err != nil || fingerprint == "" {
+	// we ignore a commonly known "product_uuid" known not to be unique
+	if err != nil || fingerprint == "" || fingerprint == "03000200-0400-0500-0006-000700080009" {
 		fingerprint = a.systemDetails.Hostname + a.systemDetails.CpuModel
 	}
 

agent/docker.go

@@ -757,7 +757,6 @@ func (dm *dockerManager) GetHostInfo() (info container.HostInfo, err error) {
 	defer resp.Body.Close()
 
 	if err := json.NewDecoder(resp.Body).Decode(&info); err != nil {
-		slog.Error("Failed to decode Docker version response", "error", err)
 		return info, err
 	}
 

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.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
+		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.Debug("NVML: Collected data", "gpu", gpu)
+	}
+}

agent/gpu_nvml_linux.go

@@ -0,0 +1,57 @@
+//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.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
+}

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
+}

agent/server_test.go

@@ -550,6 +550,9 @@ func createTestCombinedData() *system.CombinedData {
 			DiskUsed:  549755813888,  // 512GB
 			DiskPct:   50.0,
 		},
+		Details: &system.Details{
+			Hostname: "test-host",
+		},
 		Info: system.Info{
 			Uptime:       3600,
 			AgentVersion: "0.12.0",

agent/system.go

@@ -30,7 +30,7 @@ type prevDisk struct {
 }
 
 // Sets initial / non-changing values about the host system
-func (a *Agent) refreshStaticInfo() {
+func (a *Agent) refreshSystemDetails() {
 	a.systemInfo.AgentVersion = beszel.Version
 
 	// get host info from Docker if available
@@ -246,7 +246,6 @@ func (a *Agent) getSystemStats(cacheTimeMs uint16) system.Stats {
 	a.systemInfo.Uptime, _ = host.Uptime()
 	a.systemInfo.BandwidthBytes = systemStats.Bandwidth[0] + systemStats.Bandwidth[1]
 	a.systemInfo.Threads = a.systemDetails.Threads
-	slog.Debug("sysinfo", "data", a.systemInfo)
 
 	return systemStats
 }

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 v4.27.1/go.mod h1:uVtb5OGqUKpoLWhqwNQo/8LwvoiEBLvZXIQ/SmO6mL0=
-modernc.org/ccgo/v4 v4.30.1 h1:4r4U1J6Fhj98NKfSjnPUN7Ze2c6MnAdL0hWw6+LrJpc=
-modernc.org/ccgo/v4 v4.30.1/go.mod h1:bIOeI1JL54Utlxn+LwrFyjCx2n2RDiYEaJVSrgdrRfM=
+modernc.org/cc/v4 v4.26.5 h1:xM3bX7Mve6G8K8b+T11ReenJOT+BmVqQj0FY5T4+5Y4=
+modernc.org/cc/v4 v4.26.5/go.mod h1:uVtb5OGqUKpoLWhqwNQo/8LwvoiEBLvZXIQ/SmO6mL0=
+modernc.org/ccgo/v4 v4.28.1 h1:wPKYn5EC/mYTqBO373jKjvX2n+3+aK7+sICCv4Fjy1A=
+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/container/container.go

@@ -34,15 +34,12 @@ type ApiStats struct {
 	MemoryStats MemoryStats `json:"memory_stats"`
 }
 
-// Docker system info from /info
+// 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"`
-	// OSVersion       string `json:"OSVersion"`
-	// OSType          string `json:"OSType"`
-	// Architecture    string `json:"Architecture"`
 }
 
 func (s *ApiStats) CalculateCpuPercentLinux(prevCpuContainer uint64, prevCpuSystem uint64) float64 {

internal/entities/system/system.go

@@ -125,22 +125,22 @@ const (
 
 // 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
-	CpuModel      string `json:"m,omitempty" cbor:"4,keyasint,omitempty"`   // deprecated - moved to Details struct
-	Podman        bool   `json:"p,omitempty" cbor:"11,keyasint,omitempty"`  // deprecated - moved to Details struct
-	Os            Os     `json:"os,omitempty" cbor:"14,keyasint,omitempty"` // deprecated - moved to Details 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
 	Threads        int     `json:"t,omitempty" cbor:"3,keyasint,omitempty"`
+	CpuModel       string  `json:"m,omitempty" cbor:"4,keyasint,omitempty"` // deprecated - moved to Details struct
 	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
 	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
@@ -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"`
-	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"`
+	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

internal/hub/expirymap/expirymap_test.go

@@ -415,7 +415,11 @@ func TestExpiryMap_RemoveValue_WithExpiration(t *testing.T) {
 	// Wait for first value to expire
 	time.Sleep(time.Millisecond * 20)
 
-	// Try to remove the expired value - should remove one of the "value1" entries
+	// Trigger lazy cleanup of the expired key
+	_, ok := em.GetOk("key1")
+	assert.False(t, ok)
+
+	// Try to remove the remaining "value1" entry (key3)
 	removedValue, ok := em.RemovebyValue("value1")
 	assert.True(t, ok)
 	assert.Equal(t, "value1", removedValue)
@@ -423,14 +427,9 @@ func TestExpiryMap_RemoveValue_WithExpiration(t *testing.T) {
 	// Should still have key2 (different value)
 	assert.True(t, em.Has("key2"))
 
-	// 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
+	// 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"))
 }
 
 func TestExpiryMap_ValueOperations_Integration(t *testing.T) {

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 {
@@ -123,10 +124,30 @@ func (sys *System) update() error {
 	if !sys.detailsFetched.Load() {
 		options.IncludeDetails = true
 	}
+
 	data, err := sys.fetchDataFromAgent(options)
-	if err == nil {
-		_, err = sys.createRecords(data)
+	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
 }
 
@@ -197,6 +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)
@@ -208,18 +233,6 @@ func (sys *System) createRecords(data *system.CombinedData) (*core.Record, error
 		return nil
 	})
 
-	// Fetch and save SMART devices when system first comes online
-	if err == nil {
-		if !sys.smartFetched.Load() && sys.smartFetching.CompareAndSwap(false, true) {
-			go func() {
-				defer sys.smartFetching.Store(false)
-				if err := sys.FetchAndSaveSmartDevices(); err == nil {
-					sys.smartFetched.Store(true)
-				}
-			}()
-		}
-	}
-
 	return systemRecord, err
 }
 
@@ -395,8 +408,7 @@ func (sys *System) fetchStringFromAgentViaSSH(action common.WebSocketAction, req
 		if err := session.Shell(); err != nil {
 			return false, err
 		}
-		reqDataBytes, _ := cbor.Marshal(requestData)
-		req := common.HubRequest[cbor.RawMessage]{Action: action, Data: reqDataBytes}
+		req := common.HubRequest[any]{Action: action, Data: requestData}
 		_ = cbor.NewEncoder(stdin).Encode(req)
 		_ = stdin.Close()
 		var resp common.AgentResponse
@@ -460,8 +472,7 @@ func (sys *System) FetchSystemdInfoFromAgent(serviceName string) (systemd.Servic
 			return false, err
 		}
 
-		reqDataBytes, _ := cbor.Marshal(common.SystemdInfoRequest{ServiceName: serviceName})
-		req := common.HubRequest[cbor.RawMessage]{Action: common.GetSystemdInfo, Data: reqDataBytes}
+		req := common.HubRequest[any]{Action: common.GetSystemdInfo, Data: common.SystemdInfoRequest{ServiceName: serviceName}}
 		if err := cbor.NewEncoder(stdin).Encode(req); err != nil {
 			return false, err
 		}
@@ -509,8 +520,7 @@ func (sys *System) fetchDataViaSSH(options common.DataRequestOptions) (*system.C
 		*sys.data = system.CombinedData{}
 
 		if sys.agentVersion.GTE(beszel.MinVersionAgentResponse) && stdinErr == nil {
-			reqDataBytes, _ := cbor.Marshal(options)
-			req := common.HubRequest[cbor.RawMessage]{Action: common.GetData, Data: reqDataBytes}
+			req := common.HubRequest[any]{Action: common.GetData, Data: options}
 			_ = cbor.NewEncoder(stdin).Encode(req)
 			_ = stdin.Close()
 

internal/hub/systems/systems_production.go

@@ -0,0 +1,10 @@
+//go:build !testing
+// +build !testing
+
+package systems
+
+// Background SMART fetching is enabled in production but disabled for tests (systems_test_helpers.go).
+//
+// The hub integration tests create/replace systems and clean up the test apps quickly.
+// Background SMART fetching can outlive teardown and crash in PocketBase internals (nil DB).
+func backgroundSmartFetchEnabled() bool { return true }

internal/hub/systems/systems_test_helpers.go

@@ -10,6 +10,13 @@ import (
 	entities "github.com/henrygd/beszel/internal/entities/system"
 )
 
+// The hub integration tests create/replace systems and cleanup the test apps quickly.
+// Background SMART fetching can outlive teardown and crash in PocketBase internals (nil DB).
+//
+// We keep the explicit SMART refresh endpoint / method available, but disable
+// the automatic background fetch during tests.
+func backgroundSmartFetchEnabled() bool { return false }
+
 // TESTING ONLY: GetSystemCount returns the number of systems in the store
 func (sm *SystemManager) GetSystemCount() int {
 	return sm.systems.Length()

internal/site/bun.lock

@@ -36,8 +36,8 @@
         "lucide-react": "^0.452.0",
         "nanostores": "^0.11.4",
         "pocketbase": "^0.26.2",
-        "react": "^19.1.1",
-        "react-dom": "^19.1.1",
+        "react": "^19.1.2",
+        "react-dom": "^19.1.2",
         "recharts": "^2.15.4",
         "shiki": "^3.13.0",
         "tailwind-merge": "^3.3.1",
@@ -811,9 +811,9 @@
 
     "punycode": ["punycode@2.3.1", "", {}, "sha512-vYt7UD1U9Wg6138shLtLOvdAu+8DsC/ilFtEVHcH+wydcSpNE20AfSOduf6MkRFahL5FY7X1oU7nKVZFtfq8Fg=="],
 
-    "react": ["react@19.1.1", "", {}, "sha512-w8nqGImo45dmMIfljjMwOGtbmC/mk4CMYhWIicdSflH91J9TyCyczcPFXJzrZ/ZXcgGRFeP6BU0BEJTw6tZdfQ=="],
+    "react": ["react@19.2.3", "", {}, "sha512-Ku/hhYbVjOQnXDZFv2+RibmLFGwFdeeKHFcOTlrt7xplBnya5OGn/hIRDsqDiSUcfORsDC7MPxwork8jBwsIWA=="],
 
-    "react-dom": ["react-dom@19.1.1", "", { "dependencies": { "scheduler": "^0.26.0" }, "peerDependencies": { "react": "^19.1.1" } }, "sha512-Dlq/5LAZgF0Gaz6yiqZCf6VCcZs1ghAJyrsu84Q/GT0gV+mCxbfmKNoGRKBYMJ8IEdGPqu49YWXD02GCknEDkw=="],
+    "react-dom": ["react-dom@19.2.3", "", { "dependencies": { "scheduler": "^0.27.0" }, "peerDependencies": { "react": "^19.2.3" } }, "sha512-yELu4WmLPw5Mr/lmeEpox5rw3RETacE++JgHqQzd2dg+YbJuat3jH4ingc+WPZhxaoFzdv9y33G+F7Nl5O0GBg=="],
 
     "react-is": ["react-is@18.3.1", "", {}, "sha512-/LLMVyas0ljjAtoYiPqYiL8VWXzUUdThrmU5+n20DZv+a+ClRoevUzw5JxU+Ieh5/c87ytoTBV9G1FiKfNJdmg=="],
 
@@ -851,7 +851,7 @@
 
     "safe-buffer": ["safe-buffer@5.2.1", "", {}, "sha512-rp3So07KcdmmKbGvgaNxQSJr7bGVSVk5S9Eq1F+ppbRo70+YeaDxkw5Dd8NPN+GD6bjnYm2VuPuCXmpuYvmCXQ=="],
 
-    "scheduler": ["scheduler@0.26.0", "", {}, "sha512-NlHwttCI/l5gCPR3D1nNXtWABUmBwvZpEQiD4IXSbIDq8BzLIK/7Ir5gTFSGZDUu37K5cMNp0hFtzO38sC7gWA=="],
+    "scheduler": ["scheduler@0.27.0", "", {}, "sha512-eNv+WrVbKu1f3vbYJT/xtiF5syA5HPIMtf9IgY/nKg0sWqzAUEvqY/xm7OcZc/qafLx/iO9FgOmeSAp4v5ti/Q=="],
 
     "semver": ["semver@6.3.1", "", { "bin": "bin/semver.js" }, "sha512-BR7VvDCVHO+q2xBEWskxS6DJE1qRnb7DxzUrogb71CWoSficBxYsiAGd+Kl0mmq/MprG9yArRkyrQxTO6XjMzA=="],
 
@@ -971,8 +971,6 @@
 
     "prop-types/react-is": ["react-is@16.13.1", "", {}, "sha512-24e6ynE2H+OKt4kqsOvNd8kBpV65zoxbA4BVsEOB3ARVWQki/DHzaUoC5KuON/BiccDaCCTZBuOcfZs70kR8bQ=="],
 
-    "pseudolocale/commander": ["commander@10.0.1", "", {}, "sha512-y4Mg2tXshplEbSGzx7amzPwKKOCGuoSRP/CjEdwwk0FOGlUbq6lKuoyDZTNZkmxHdJtp54hdfY/JUrdL7Xfdug=="],
-
     "readdirp/picomatch": ["picomatch@2.3.1", "", {}, "sha512-JU3teHTNjmE2VCGFzuY8EXzCDVwEqB2a8fsIvwaStHhAWJEeVd1o1QD80CU6+ZdEXXSLbSsuLwJjkCBWqRQUVA=="],
 
     "restore-cursor/signal-exit": ["signal-exit@3.0.7", "", {}, "sha512-wnD2ZE+l+SPC/uoS0vXeE9L1+0wuaMqKlfz9AMUo38JsyLSBWSFcHR1Rri62LZc12vLr1gb3jl7iwQhgwpAbGQ=="],
@@ -981,28 +979,18 @@
 
     "string-width-cjs/emoji-regex": ["emoji-regex@8.0.0", "", {}, "sha512-MSjYzcWNOA0ewAHpz0MxpYFvwg6yjy1NG3xteoqz644VCo/RPgnr1/GGt+ic3iJTzQ8Eu3TdM14SawnVUmGE6A=="],
 
-    "string-width-cjs/strip-ansi": ["strip-ansi@6.0.1", "", { "dependencies": { "ansi-regex": "^5.0.1" } }, "sha512-Y38VPSHcqkFrCpFnQ9vuSXmquuv5oXOKpGeT6aGrr3o3Gc9AlVa6JBfUSOCnbxGGZF+/0ooI7KrPuUSztUdU5A=="],
-
-    "strip-ansi-cjs/ansi-regex": ["ansi-regex@5.0.1", "", {}, "sha512-quJQXlTSUGL2LH9SUXo8VwsY4soanhgo6LNSm84E1LBcE8s3O0wpdiRzyR9z/ZZJMlMWv37qOOb9pdJlMUEKFQ=="],
-
     "wrap-ansi/ansi-styles": ["ansi-styles@6.2.1", "", {}, "sha512-bN798gFfQX+viw3R7yrGWRqnrN2oRkEkUjjl4JNn4E8GxxbjtG3FbrEIIY3l8/hrwUwIeCZvi4QuOTP4MErVug=="],
 
     "wrap-ansi/strip-ansi": ["strip-ansi@7.1.0", "", { "dependencies": { "ansi-regex": "^6.0.1" } }, "sha512-iq6eVVI64nQQTRYq2KtEg2d2uU7LElhTJwsH4YzIHZshxlgZms/wIc4VoDQTlG/IvVIrBKG06CrZnp0qv7hkcQ=="],
 
     "wrap-ansi-cjs/string-width": ["string-width@4.2.3", "", { "dependencies": { "emoji-regex": "^8.0.0", "is-fullwidth-code-point": "^3.0.0", "strip-ansi": "^6.0.1" } }, "sha512-wKyQRQpjJ0sIp62ErSZdGsjMJWsap5oRNihHhu6G7JVO/9jIB6UyevL+tXuOqrng8j/cxKTWyWUwvSTriiZz/g=="],
 
-    "wrap-ansi-cjs/strip-ansi": ["strip-ansi@6.0.1", "", { "dependencies": { "ansi-regex": "^5.0.1" } }, "sha512-Y38VPSHcqkFrCpFnQ9vuSXmquuv5oXOKpGeT6aGrr3o3Gc9AlVa6JBfUSOCnbxGGZF+/0ooI7KrPuUSztUdU5A=="],
-
     "@isaacs/cliui/strip-ansi/ansi-regex": ["ansi-regex@6.0.1", "", {}, "sha512-n5M855fKb2SsfMIiFFoVrABHJC8QtHwVx+mHWP3QcEqBHYienj5dHSgjbxtC0WEZXYt4wcD6zrQElDPhFuZgfA=="],
 
-    "string-width-cjs/strip-ansi/ansi-regex": ["ansi-regex@5.0.1", "", {}, "sha512-quJQXlTSUGL2LH9SUXo8VwsY4soanhgo6LNSm84E1LBcE8s3O0wpdiRzyR9z/ZZJMlMWv37qOOb9pdJlMUEKFQ=="],
-
     "string-width/strip-ansi/ansi-regex": ["ansi-regex@6.0.1", "", {}, "sha512-n5M855fKb2SsfMIiFFoVrABHJC8QtHwVx+mHWP3QcEqBHYienj5dHSgjbxtC0WEZXYt4wcD6zrQElDPhFuZgfA=="],
 
     "wrap-ansi-cjs/string-width/emoji-regex": ["emoji-regex@8.0.0", "", {}, "sha512-MSjYzcWNOA0ewAHpz0MxpYFvwg6yjy1NG3xteoqz644VCo/RPgnr1/GGt+ic3iJTzQ8Eu3TdM14SawnVUmGE6A=="],
 
-    "wrap-ansi-cjs/strip-ansi/ansi-regex": ["ansi-regex@5.0.1", "", {}, "sha512-quJQXlTSUGL2LH9SUXo8VwsY4soanhgo6LNSm84E1LBcE8s3O0wpdiRzyR9z/ZZJMlMWv37qOOb9pdJlMUEKFQ=="],
-
     "wrap-ansi/strip-ansi/ansi-regex": ["ansi-regex@6.0.1", "", {}, "sha512-n5M855fKb2SsfMIiFFoVrABHJC8QtHwVx+mHWP3QcEqBHYienj5dHSgjbxtC0WEZXYt4wcD6zrQElDPhFuZgfA=="],
   }
 }

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

@@ -46,6 +46,7 @@ import type {
 	ChartTimes,
 	ContainerStatsRecord,
 	GPUData,
+	SystemDetailsRecord,
 	SystemInfo,
 	SystemRecord,
 	SystemStats,
@@ -166,7 +167,7 @@ export default memo(function SystemDetail({ id }: { id: string }) {
 	const isLongerChart = !["1m", "1h"].includes(chartTime) // true if chart time is not 1m or 1h
 	const userSettings = $userSettings.get()
 	const chartWrapRef = useRef<HTMLDivElement>(null)
-	const [isPodman, setIsPodman] = useState(system.info?.p ?? false)
+	const [details, setDetails] = useState<SystemDetailsRecord>({} as SystemDetailsRecord)
 
 	useEffect(() => {
 		return () => {
@@ -176,6 +177,7 @@ export default memo(function SystemDetail({ id }: { id: string }) {
 			persistChartTime.current = false
 			setSystemStats([])
 			setContainerData([])
+			setDetails({} as SystemDetailsRecord)
 			$containerFilter.set("")
 		}
 	}, [id])
@@ -203,6 +205,22 @@ export default memo(function SystemDetail({ id }: { id: string }) {
 		}
 	}, [system?.info?.v])
 
+	// fetch system details
+	useEffect(() => {
+		// if system.info.m exists, agent is old version without system details
+		if (!system.id || system.info?.m) {
+			return
+		}
+		pb.collection<SystemDetailsRecord>("system_details")
+			.getOne(system.id, {
+				fields: "hostname,kernel,cores,threads,cpu,os,os_name,arch,memory,podman",
+				headers: {
+					"Cache-Control": "public, max-age=60",
+				},
+			})
+			.then(setDetails)
+	}, [system.id])
+
 	// subscribe to realtime metrics if chart time is 1m
 	// biome-ignore lint/correctness/useExhaustiveDependencies: not necessary
 	useEffect(() => {
@@ -322,10 +340,6 @@ export default memo(function SystemDetail({ id }: { id: string }) {
 		})
 	}, [system, chartTime])
 
-	useEffect(() => {
-		setIsPodman(system.info?.p ?? false)
-	}, [system.info?.p])
-
 	/** Space for tooltip if more than 10 sensors and no containers table */
 	useEffect(() => {
 		const sensors = Object.keys(systemStats.at(-1)?.stats.t ?? {})
@@ -393,12 +407,14 @@ export default memo(function SystemDetail({ id }: { id: string }) {
 	const hasGpuData = lastGpuVals.length > 0
 	const hasGpuPowerData = lastGpuVals.some((gpu) => gpu.p !== undefined || gpu.pp !== undefined)
 	const hasGpuEnginesData = lastGpuVals.some((gpu) => gpu.e !== undefined)
+	const isLinux = (details?.os ?? system.info?.os) === Os.Linux
+	const isPodman = details?.podman ?? system.info?.p ?? false
 
 	return (
 		<>
 			<div ref={chartWrapRef} className="grid gap-4 mb-14 overflow-x-clip">
 				{/* system info */}
-				<InfoBar system={system} chartData={chartData} grid={grid} setGrid={setGrid} setIsPodman={setIsPodman} />
+				<InfoBar system={system} chartData={chartData} grid={grid} setGrid={setGrid} details={details} />
 
 				{/* <Tabs defaultValue="overview" className="w-full">
 					<TabsList className="w-full h-11">
@@ -841,7 +857,7 @@ export default memo(function SystemDetail({ id }: { id: string }) {
 					<LazyContainersTable systemId={system.id} />
 				)}
 
-				{system.info?.os === Os.Linux && compareSemVer(chartData.agentVersion, parseSemVer("0.16.0")) >= 0 && (
+				{isLinux && compareSemVer(chartData.agentVersion, parseSemVer("0.16.0")) >= 0 && (
 					<LazySystemdTable systemId={system.id} />
 				)}
 			</div>

internal/site/src/components/routes/system/info-bar.tsx

@@ -11,14 +11,13 @@ import {
 	MonitorIcon,
 	Rows,
 } from "lucide-react"
-import { useEffect, useMemo, useState } from "react"
+import { useMemo } from "react"
 import ChartTimeSelect from "@/components/charts/chart-time-select"
 import { Button } from "@/components/ui/button"
 import { Card } from "@/components/ui/card"
 import { FreeBsdIcon, TuxIcon, WebSocketIcon, WindowsIcon } from "@/components/ui/icons"
 import { Separator } from "@/components/ui/separator"
 import { Tooltip, TooltipContent, TooltipProvider, TooltipTrigger } from "@/components/ui/tooltip"
-import { pb } from "@/lib/api"
 import { ConnectionType, connectionTypeLabels, Os, SystemStatus } from "@/lib/enums"
 import { cn, formatBytes, getHostDisplayValue, secondsToString, toFixedFloat } from "@/lib/utils"
 import type { ChartData, SystemDetailsRecord, SystemRecord } from "@/types"
@@ -28,44 +27,15 @@ export default function InfoBar({
 	chartData,
 	grid,
 	setGrid,
-	setIsPodman,
+	details,
 }: {
 	system: SystemRecord
 	chartData: ChartData
 	grid: boolean
 	setGrid: (grid: boolean) => void
-	setIsPodman: (isPodman: boolean) => void
+	details: SystemDetailsRecord | null
 }) {
 	const { t } = useLingui()
-	const [details, setDetails] = useState<SystemDetailsRecord | null>(null)
-
-	// Fetch system_details on mount / when system changes
-	useEffect(() => {
-		let active = true
-		setDetails(null)
-		// skip fetching system details if agent is older version which includes details in Info struct
-		if (!system.id || system.info?.m) {
-			return
-		}
-		pb.collection<SystemDetailsRecord>("system_details")
-			.getOne(system.id, {
-				fields: "hostname,kernel,cores,threads,cpu,os,os_name,arch,memory,podman",
-				headers: {
-					"Cache-Control": "public, max-age=60",
-				},
-			})
-			.then((details) => {
-				if (active) {
-					setDetails(details)
-					setIsPodman(details.podman)
-				}
-			})
-			.catch(() => {})
-
-		return () => {
-			active = false
-		}
-	}, [system.id])
 
 	// values for system info bar - use details with fallback to system.info
 	const systemInfo = useMemo(() => {

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",
-		sortingFn: (a, b) => a.original.device.localeCompare(b.original.device),
+		accessorKey: "name",
+		sortingFn: (a, b) => a.original.name.localeCompare(b.original.name),
 		header: ({ column }) => <HeaderButton column={column} name={t`Device`} Icon={HardDrive} />,
-		cell: ({ row }) => (
-			<div className="font-medium max-w-40 truncate ms-1.5" title={row.getValue("device")}>
-				{row.getValue("device")}
+		cell: ({ getValue }) => (
+			<div className="font-medium max-w-40 truncate ms-1.5" title={getValue() as string}>
+				{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 }) => (
-			<div className="max-w-48 truncate ms-1.5" title={row.getValue("model")}>
-				{row.getValue("model")}
+		cell: ({ getValue }) => (
+			<div className="max-w-48 truncate ms-1.5" title={getValue() as string}>
+				{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",
-		sortingFn: (a, b) => a.original.deviceType.localeCompare(b.original.deviceType),
+		accessorKey: "type",
+		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",
-	// 	sortingFn: (a, b) => a.original.serialNumber.localeCompare(b.original.serialNumber),
+	// 	accessorKey: "serial",
+	// 	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",
-	// 	sortingFn: (a, b) => a.original.firmwareVersion.localeCompare(b.original.firmwareVersion),
+	// 	accessorKey: "firmware",
+	// 	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)
-	const fetchSmartDevices = useCallback(() => {
+	// Fetch smart devices
+	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) => {
-				setSmartDevices(records)
+			.then(setSmartDevices)
+			.catch((err) => {
+				if (!err.isAbort) {
+					setSmartDevices([])
+				}
 			})
-			.catch(() => setSmartDevices([]))
+
+		return () => controller.abort()
 	}, [systemId])
 
-	// Fetch smart devices when component mounts or systemId changes
-	useEffect(() => {
-		fetchSmartDevices()
-	}, [fetchSmartDevices])
-
-	// Subscribe to live updates so rows add/remove without manual refresh/filtering
+	// Subscribe to updates
 	useEffect(() => {
 		let unsubscribe: (() => void) | undefined
 		const pbOptions = systemId
 			? { fields: SMART_DEVICE_FIELDS, filter: pb.filter("system = {:system}", { system: systemId }) }
 			: { fields: SMART_DEVICE_FIELDS }
 
-		;(async () => {
-			try {
-				unsubscribe = await pb.collection("smart_devices").subscribe(
-					"*",
-					(event) => {
-						const record = event.record as SmartDeviceRecord
-						setSmartDevices((currentDevices) => {
-							const devices = currentDevices ?? []
-							const matchesSystemScope = !systemId || record.system === systemId
+			; (async () => {
+				try {
+					unsubscribe = await pb.collection("smart_devices").subscribe(
+						"*",
+						(event) => {
+							const record = event.record as SmartDeviceRecord
+							setSmartDevices((currentDevices) => {
+								const devices = currentDevices ?? []
+								const matchesSystemScope = !systemId || record.system === systemId
 
-							if (event.action === "delete") {
-								return devices.filter((device) => device.id !== record.id)
-							}
+								if (event.action === "delete") {
+									return devices.filter((device) => device.id !== record.id)
+								}
 
-							if (!matchesSystemScope) {
-								// Record moved out of scope; ensure it disappears locally.
-								return devices.filter((device) => device.id !== record.id)
-							}
+								if (!matchesSystemScope) {
+									// Record moved out of scope; ensure it disappears locally.
+									return devices.filter((device) => device.id !== record.id)
+								}
 
-							const existingIndex = devices.findIndex((device) => device.id === record.id)
-							if (existingIndex === -1) {
-								return [record, ...devices]
-							}
+								const existingIndex = devices.findIndex((device) => device.id === record.id)
+								if (existingIndex === -1) {
+									return [record, ...devices]
+								}
 
-							const next = [...devices]
-							next[existingIndex] = record
-							return next
-						})
-					},
-					pbOptions
-				)
-			} catch (error) {
-				console.error("Failed to subscribe to SMART device updates:", error)
-			}
-		})()
+								const next = [...devices]
+								next[existingIndex] = record
+								return next
+							})
+						},
+						pbOptions
+					)
+				} catch (error) {
+					console.error("Failed to subscribe to SMART device updates:", error)
+				}
+			})()
 
 		return () => {
 			unsubscribe?.()
 		}
 	}, [systemId])
 
-	const handleRowRefresh = useCallback(
-		async (disk: DiskInfo) => {
-			if (!disk.system) return
-			setRowActionState({ type: "refresh", id: disk.id })
-			try {
-				await pb.send("/api/beszel/smart/refresh", {
-					method: "POST",
-					query: { system: disk.system },
-				})
-			} catch (error) {
-				console.error("Failed to refresh SMART device:", error)
-			} finally {
-				setRowActionState((state) => (state?.id === disk.id ? null : state))
-			}
-		},
-		[fetchSmartDevices]
-	)
+	const handleRowRefresh = useCallback(async (disk: SmartDeviceRecord) => {
+		if (!disk.system) return
+		setRowActionState({ type: "refresh", id: disk.id })
+		try {
+			await pb.send("/api/beszel/smart/refresh", {
+				method: "POST",
+				query: { system: disk.system },
+			})
+		} catch (error) {
+			console.error("Failed to refresh SMART device:", error)
+		} finally {
+			setRowActionState((state) => (state?.id === disk.id ? null : state))
+		}
+	}, [])
 
-	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 = $allSystemsById.get()[disk.system]?.name ?? ""
-			const device = disk.device ?? ""
+			const device = disk.name ?? ""
 			const model = disk.model ?? ""
-			const status = disk.status ?? ""
-			const type = disk.deviceType ?? ""
+			const status = disk.state ?? ""
+			const type = disk.type ?? ""
 			const searchString = `${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

@@ -1,8 +1,14 @@
 ## 0.18.0
 
-- Remove `la1`, `la5`, `la15` fields from `Info` struct in favor of `la` array.
+- Collect S.M.A.R.T. data in the background every hour.
 
-- Remove `MB` bandwidth values in favor of bytes.
+- 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