Real-time monitoring dashboard for Proxmox VE, PBS, Docker, Kubernetes, TrueNAS and vSphere. Self-hosted, with alerts for failed backups, full disks, unhealthy containers and offline nodes.
Go aialertsdashboarddockerdocker-monitoringgohomelabinfrastructure-monitoringkubernetesmonitoringmspproxmoxproxmox-backup-serverproxmox-mail-gatewayproxmox-veself-hostedtruenastypescriptvspherewebhooks
Keep absent host disk readings out of live charts and history
pulse-triage[bot]c93757fb4b11
diff --git a/docs/release-control/v6/internal/subsystems/agent-lifecycle.md b/docs/release-control/v6/internal/subsystems/agent-lifecycle.mdindex e031fa86c7..e655dc19d1 100644--- a/docs/release-control/v6/internal/subsystems/agent-lifecycle.md+++ b/docs/release-control/v6/internal/subsystems/agent-lifecycle.md@@ -15,6 +15,16 @@## Purpose+### Accepted host reports preserve disk-reading presence++An accepted host report is not evidence of a disk measurement. Disk history+recording and live fallback share canonical first-filesystem admission:+unavailable/absent readings stay absent while measured zero remains available.+Enrollment, receipt freshness, removal, correlation and stored observations keep+their existing semantics. The host disk recording and API presence regressions+cover this distinction; they do not establish installed guest recovery.++### Command access is a named token-step choice, not a legacy option (21 October 2026)The installer's command-access checkbox reads **Allow Pulse to run commands ondiff --git a/docs/release-control/v6/internal/subsystems/api-contracts.md b/docs/release-control/v6/internal/subsystems/api-contracts.mdindex 86758f19ed..68cbdea54e 100644--- a/docs/release-control/v6/internal/subsystems/api-contracts.md+++ b/docs/release-control/v6/internal/subsystems/api-contracts.md@@ -21,6 +21,22 @@## Purpose+### Host-agent disk history distinguishes absence from zero++History and full/infrastructure summary charts synthesize host-agent disk usage+only when its canonical disk reading is present. An accepted report without a+filesystem reading must not create a live 0% sample; a measured 0% remains+available. A recorded disk series keeps its original values and timestamps,+including when CPU has no history. Existing lease, error, query-selection and+provider-aggregate contracts remain unchanged.++`TestHostDiskUsageHistoryAndSummaryRespectPresence` exercises cold and recorded+History (all/selected metrics), both summary routes and disk-only selection for+missing, unavailable, measured-zero, low-percentage and nonzero readings on+online/offline hosts. Browser proof must exercise the returned producer payload+through the existing Machines drawer at desktop and narrow widths.++### Webhook deletion has an explicit committed acknowledgement`DELETE /api/notifications/webhooks/{id}` returns `success: true` after saved anddiff --git a/docs/release-control/v6/internal/subsystems/monitoring.md b/docs/release-control/v6/internal/subsystems/monitoring.mdindex 3b9abdd4b6..f3babc1da0 100644--- a/docs/release-control/v6/internal/subsystems/monitoring.md+++ b/docs/release-control/v6/internal/subsystems/monitoring.md@@ -17,6 +17,17 @@## Purpose+### Missing host-agent disk readings are not recorded as idle++`ApplyHostReport` applies the canonical first-filesystem disk admission before+writing host disk usage to recent memory or the persistent metrics store.+Absent/unavailable inventory emits no disk sample; an observed zero still emits+zero. Other accepted host fields and existing observations are not withdrawn,+rewritten or deleted. No guest operation or collection/retry policy changes.+`TestHostDiskUsageRecordingSkipsAbsenceAndPreservesZero` covers accepted absent,+zero and absent-again reports, retained history and ordinary store reopen.++### Ended backup jobs do not supply missing guest verdictsFor multi-guest vzdump jobs, only an explicit guest finish or failure linediff --git a/docs/release-control/v6/internal/subsystems/unified-resources.md b/docs/release-control/v6/internal/subsystems/unified-resources.mdindex b9004a204e..1faf910f76 100644--- a/docs/release-control/v6/internal/subsystems/unified-resources.md+++ b/docs/release-control/v6/internal/subsystems/unified-resources.md@@ -23,6 +23,18 @@ and sort the complete canonical change table while startup and ingestion wait.## Purpose+### Host disk usage retains explicit presence++`HostDiskUsagePercent` shares the existing canonical first-filesystem admission+with host metric projection, history recording and legacy live fallback.+`HostView.DiskUsagePercent` returns the presence of the canonical disk metric,+not the zero default of a missing scalar accessor. Provider aggregate readings+need no agent filesystem inventory. Neither helper establishes freshness or+changes selection/identity. `TestHostDiskUsagePercentMatchesCanonicalPresence`+pins absent/invalid/zero/low readings, unchanged first-filesystem selection and+provider aggregate zero.++### Drawer history follows current source authorityResource drawers withdraw current chart fallbacks and cached live-only repliesdiff --git a/internal/api/chartapi/service.go b/internal/api/chartapi/service.goindex 662c65648f..6d507582cf 100644--- a/internal/api/chartapi/service.go+++ b/internal/api/chartapi/service.go@@ -362,10 +362,18 @@ func (r *Service) HandleCharts(w http.ResponseWriter, req *http.Request) {if batchMetrics, ok := agentBatchMetrics[request.SQLResourceID]; ok {oldestTimestamp = fillChartSeriesFromBatch(agentData[hID], batchMetrics, oldestTimestamp)}- if len(agentData[hID]["cpu"]) == 0 && hostAgentCanSupplyLiveChartFallback(h) {- agentData[hID]["cpu"] = []MetricPoint{{Timestamp: currentTime, Value: h.CPUPercent()}}- agentData[hID]["memory"] = []MetricPoint{{Timestamp: currentTime, Value: h.MemoryPercent()}}- agentData[hID]["disk"] = []MetricPoint{{Timestamp: currentTime, Value: h.DiskPercent()}}+ if hostAgentCanSupplyLiveChartFallback(h) {+ if len(agentData[hID]["cpu"]) == 0 {+ agentData[hID]["cpu"] = []MetricPoint{{Timestamp: currentTime, Value: h.CPUPercent()}}+ }+ if len(agentData[hID]["memory"]) == 0 {+ agentData[hID]["memory"] = []MetricPoint{{Timestamp: currentTime, Value: h.MemoryPercent()}}+ }+ if len(agentData[hID]["disk"]) == 0 {+ if value, ok := h.DiskUsagePercent(); ok {+ agentData[hID]["disk"] = []MetricPoint{{Timestamp: currentTime, Value: value}}+ }+ }}}@@ -1997,7 +2005,7 @@ func (r *Service) HandleInfrastructureCharts(w http.ResponseWriter, req *http.Recase "memory":value = h.MemoryPercent()case "disk":- value = h.DiskPercent()+ value, hasFallbackValue = h.DiskUsagePercent()default:hasFallbackValue = false}diff --git a/internal/api/chartapi/service_handlers_test.go b/internal/api/chartapi/service_handlers_test.goindex 27c57f422a..e360f0a076 100644--- a/internal/api/chartapi/service_handlers_test.go+++ b/internal/api/chartapi/service_handlers_test.go@@ -206,7 +206,7 @@ func TestHandleCharts_UsesCanonicalMockUnifiedReadStateForVMwareHosts(t *testingHostname: "live-store-host-1",CPUUsage: 11.0,Memory: models.Memory{Usage: 22.0},- Disks: []models.Disk{{Usage: 33.0}},+ Disks: []models.Disk{{Total: 100, Used: 33, Usage: 33.0}},Status: "online",}}syncTestResourceStore(t, monitor, state)@@ -310,7 +310,7 @@ func TestHandleInfrastructureCharts_Lightweight(t *testing.T) {Hostname: "host-one",CPUUsage: 11.0,Memory: models.Memory{Usage: 22.0},- Disks: []models.Disk{{Usage: 33.0}},+ Disks: []models.Disk{{Total: 100, Used: 33, Usage: 33.0}},Status: "online",}}syncTestResourceStore(t, monitor, state)@@ -393,7 +393,7 @@ func TestHandleInfrastructureCharts_MetricFilter(t *testing.T) {Hostname: "host-one",CPUUsage: 11.0,Memory: models.Memory{Usage: 22.0},- Disks: []models.Disk{{Usage: 33.0}},+ Disks: []models.Disk{{Total: 100, Used: 33, Usage: 33.0}},Status: "online",}}syncTestResourceStore(t, monitor, state)@@ -448,7 +448,7 @@ func TestHandleInfrastructureCharts_UsesCanonicalMockUnifiedReadStateForVMwareHoHostname: "live-store-host-1",CPUUsage: 11.0,Memory: models.Memory{Usage: 22.0},- Disks: []models.Disk{{Usage: 33.0}},+ Disks: []models.Disk{{Total: 100, Used: 33, Usage: 33.0}},Status: "online",}}syncTestResourceStore(t, monitor, state)diff --git a/internal/api/contract_test.go b/internal/api/contract_test.goindex 1a70e11215..0a40eed941 100644--- a/internal/api/contract_test.go+++ b/internal/api/contract_test.go@@ -27873,3 +27873,90 @@ func TestContract_ChartIndexFallbackPreservesObservations(t *testing.T) {}}}++func TestHostDiskUsageHistoryAndSummaryRespectPresence(t *testing.T) {+ setMockModeForTest(t, false)+ for _, tc := range []struct {+ name string+ disks []models.Disk+ present bool+ value float64+ }{+ {name: "absent"},+ {name: "empty", disks: []models.Disk{}},+ {name: "capacity unavailable", disks: []models.Disk{{Usage: 0}}},+ {name: "unavailable", disks: []models.Disk{{Total: 100, Usage: -1}}},+ {name: "measured zero", disks: []models.Disk{{Total: 100, Free: 100, Usage: 0}}, present: true},+ {name: "small percent", disks: []models.Disk{{Total: 1000, Used: 2, Usage: 0.2}}, present: true, value: 0.2},+ {name: "measured usage", disks: []models.Disk{{Total: 100, Used: 61, Usage: 61}}, present: true, value: 61},+ } {+ t.Run(tc.name, func(t *testing.T) {+ for _, stored := range []bool{false, true} {+ for _, status := range []string{"online", "offline"} {+ t.Run(status+map[bool]string{false: "/cold", true: "/recorded"}[stored], func(t *testing.T) {+ monitor, state, history := newTestMonitor(t)+ state.Hosts = []models.Host{{ID: "fixture-host", Hostname: "Disk presence host", Status: status, LastSeen: time.Now(), CPUUsage: 12, Memory: models.Memory{Total: 8192, Used: 2048, Usage: 25}, Disks: tc.disks}}+ syncTestResourceStore(t, monitor, state)+ recordedAt := time.Now().Add(-2 * time.Minute)+ if stored {+ history.AddGuestMetric("agent:fixture-host", "disk", 43, recordedAt)+ }+ router := &Router{monitor: monitor}+ for _, path := range []string{"/api/metrics-store/history?resourceType=agent&resourceId=fixture-host&range=1h", "/api/metrics-store/history?resourceType=agent&resourceId=fixture-host&metric=disk&range=1h", "/api/charts?range=1h", "/api/charts/infrastructure?range=1h", "/api/charts/infrastructure?range=1h&metrics=disk"} {+ rec := httptest.NewRecorder()+ req := httptest.NewRequest(http.MethodGet, path, nil)+ switch {+ case strings.HasPrefix(path, "/api/metrics-store/history"):+ router.handleMetricsHistory(rec, req)+ case strings.HasPrefix(path, "/api/charts/infrastructure"):+ router.handleInfrastructureCharts(rec, req)+ default:+ router.handleCharts(rec, req)+ }+ want := stored || (status == "online" && tc.present)+ var payload struct {+ Points []chartapi.MetricPoint `json:"points"`+ Metrics map[string][]chartapi.MetricPoint `json:"metrics"`+ AgentData map[string]map[string][]chartapi.MetricPoint `json:"agentData"`+ }+ // A wholly unmeasured history target keeps its existing no-source error;+ // a missing metric must never turn into a successful zero sample.+ if rec.Code == http.StatusServiceUnavailable && !want && strings.HasPrefix(path, "/api/metrics-store/history") {+ continue+ }+ if rec.Code != http.StatusOK {+ t.Fatalf("%s status %d: %s", path, rec.Code, rec.Body.String())+ }+ if err := json.Unmarshal(rec.Body.Bytes(), &payload); err != nil {+ t.Fatal(err)+ }+ points := payload.Points+ if !strings.Contains(path, "metric=disk") {+ if payload.AgentData != nil {+ points = payload.AgentData["fixture-host"]["disk"]+ } else {+ points = payload.Metrics["disk"]+ }+ }+ if (len(points) > 0) != want {+ t.Fatalf("%s disk presence=%v want=%v; %s", path, len(points) > 0, want, rec.Body.String())+ }+ if want {+ value := tc.value+ if stored {+ value = 43+ }+ if len(points) != 1 || points[0].Value != value {+ t.Fatalf("%s lost real reading: %+v want %v", path, points, value)+ }+ if stored && points[0].Timestamp != recordedAt.UnixMilli() {+ t.Fatalf("%s replaced original recorded timestamp: %+v", path, points)+ }+ }+ }+ })+ }+ }+ })+ }+}diff --git a/internal/api/router.go b/internal/api/router.goindex 05c9663f75..d92ff447d3 100644--- a/internal/api/router.go+++ b/internal/api/router.go@@ -6713,11 +6713,9 @@ func (r *Router) handleMetricsHistory(w http.ResponseWriter, req *http.Request)if host != nil {points["cpu"] = monitoring.MetricPoint{Timestamp: now, Value: host.CPUUsage}points["memory"] = monitoring.MetricPoint{Timestamp: now, Value: host.Memory.Usage}- diskPercent := float64(0)- if len(host.Disks) > 0 {- diskPercent = host.Disks[0].Usage+ if diskPercent, ok := unifiedresources.HostDiskUsagePercent(host.Disks); ok {+ points["disk"] = monitoring.MetricPoint{Timestamp: now, Value: diskPercent}}- points["disk"] = monitoring.MetricPoint{Timestamp: now, Value: diskPercent}if temperature := models.HostSensorPrimaryCPUTemperatureCelsius(host.Sensors); temperature != nil {points["temperature"] = monitoring.MetricPoint{Timestamp: now, Value: *temperature}}diff --git a/internal/monitoring/monitor_agents.go b/internal/monitoring/monitor_agents.goindex 6305f771b7..a5df94fa92 100644--- a/internal/monitoring/monitor_agents.go+++ b/internal/monitoring/monitor_agents.go@@ -3837,10 +3837,7 @@ func (m *Monitor) ApplyHostReport(report agentshost.Report, tokenRecord *config.// Record host-agent metrics for sparkline charts.hostMetricKey := fmt.Sprintf("agent:%s", host.ID)- var hostDiskPercent float64- if len(host.Disks) > 0 {- hostDiskPercent = host.Disks[0].Usage- }+ hostDiskPercent, hasHostDiskUsage := unifiedresources.HostDiskUsagePercent(host.Disks)hostTemperature := models.HostSensorPrimaryCPUTemperatureCelsius(host.Sensors)hostGPUUtilization, hostGPUMemory, hostGPUTemperature := hostGPUMetrics(host.Sensors)@@ -3850,7 +3847,9 @@ func (m *Monitor) ApplyHostReport(report agentshost.Report, tokenRecord *config.if host.Memory.HasKnownUsage() {m.metricsHistory.AddGuestMetric(hostMetricKey, "memory", host.Memory.Usage, now)}- m.metricsHistory.AddGuestMetric(hostMetricKey, "disk", hostDiskPercent, now)+ if hasHostDiskUsage {+ m.metricsHistory.AddGuestMetric(hostMetricKey, "disk", hostDiskPercent, now)+ }if hostTemperature != nil {m.metricsHistory.AddGuestMetric(hostMetricKey, "temperature", *hostTemperature, now)}@@ -3885,7 +3884,9 @@ func (m *Monitor) ApplyHostReport(report agentshost.Report, tokenRecord *config.if host.Memory.HasKnownUsage() {m.metricsStore.Write("agent", host.ID, "memory", host.Memory.Usage, now)}- m.metricsStore.Write("agent", host.ID, "disk", hostDiskPercent, now)+ if hasHostDiskUsage {+ m.metricsStore.Write("agent", host.ID, "disk", hostDiskPercent, now)+ }if hostTemperature != nil {m.metricsStore.Write("agent", host.ID, "temperature", *hostTemperature, now)}diff --git a/internal/monitoring/monitor_host_agents_test.go b/internal/monitoring/monitor_host_agents_test.goindex 31fd41d691..fc63622366 100644--- a/internal/monitoring/monitor_host_agents_test.go+++ b/internal/monitoring/monitor_host_agents_test.go@@ -8753,3 +8753,75 @@ func TestMonitorDiscoveryEventsStayWithOwningOrganization(t *testing.T) {}}}++func TestHostDiskUsageRecordingSkipsAbsenceAndPreservesZero(t *testing.T) {+ cfg := metrics.DefaultConfig(t.TempDir())+ cfg.WriteBufferSize = 1+ store, err := metrics.NewStore(cfg)+ if err != nil {+ t.Fatal(err)+ }+ closed := false+ t.Cleanup(func() {+ if !closed {+ store.Close()+ }+ })+ monitor := issue1654Monitor()+ monitor.metricsHistory = NewMetricsHistory(20, time.Hour)+ monitor.metricsStore = store+ // Close drains ordinary queued writes before the final read; no polling for+ // a favourable absence result and no changes to the store or its driver.+ report := issue1654Report(time.Now().UTC())+ report.Disks = nil+ report.Metrics = agentshost.Metrics{CPUUsagePercent: 12, Memory: agentshost.MemoryMetric{TotalBytes: 8192, UsedBytes: 2048, Usage: 25}}+ report.SequenceID = agentshost.FormatReportSequenceID("disk-presence", 1)+ host, err := monitor.ApplyHostReport(report, nil)+ if err != nil {+ t.Fatal(err)+ }+ if points := monitor.metricsHistory.GetGuestMetrics("agent:"+host.ID, "disk", time.Hour); len(points) != 0 {+ t.Fatalf("absent inventory recorded disk zero: %+v", points)+ }+ recordedAt := time.Now().Add(-2 * time.Minute)+ monitor.metricsHistory.AddGuestMetric("agent:"+host.ID, "disk", 43, recordedAt)+ store.Write("agent", host.ID, "disk", 43, recordedAt)+ report.Disks = []agentshost.Disk{{Device: "/dev/sda", Mountpoint: "/", Type: "ext4", TotalBytes: 100, FreeBytes: 100, Usage: 0}}+ report.SequenceID = agentshost.FormatReportSequenceID("disk-presence", 2)+ report.Timestamp = time.Now().UTC()+ if _, err := monitor.ApplyHostReport(report, nil); err != nil {+ t.Fatal(err)+ }+ report.Disks = nil+ report.SequenceID = agentshost.FormatReportSequenceID("disk-presence", 3)+ report.Timestamp = time.Now().UTC()+ if _, err := monitor.ApplyHostReport(report, nil); err != nil {+ t.Fatal(err)+ }+ points := monitor.metricsHistory.GetGuestMetrics("agent:"+host.ID, "disk", time.Hour)+ if len(points) != 2 || points[0].Value != 43 || points[1].Value != 0 || !points[0].Timestamp.Equal(recordedAt) {+ t.Fatalf("absence changed recorded/real-zero observations: %+v", points)+ }+ if err := store.Close(); err != nil {+ t.Fatal(err)+ }+ closed = true+ reopened, err := metrics.NewStore(cfg)+ if err != nil {+ t.Fatal(err)+ }+ defer reopened.Close()+ saved, err := reopened.Query("agent", host.ID, "disk", recordedAt.Add(-time.Minute), time.Now().Add(time.Minute), 0)+ if err != nil {+ t.Fatal(err)+ }+ if len(saved) != 2 || saved[0].Value != 43 || saved[1].Value != 0 {+ t.Fatalf("persisted disk series invented or lost a sample: %+v", saved)+ }+ if saved[0].Timestamp.UnixMilli() != recordedAt.UnixMilli() {+ t.Fatalf("persisted observation time changed: %+v", saved[0])+ }+ if got := monitor.HostsSnapshot(); len(got) != 1 || len(got[0].Disks) != 0 || got[0].Memory.Total != 8192 {+ t.Fatalf("accepted no-disk report was not retained: %+v", got)+ }+}diff --git a/internal/unifiedresources/host_disk_usage.go b/internal/unifiedresources/host_disk_usage.gonew file mode 100644index 0000000000..3bfe93ba4f--- /dev/null+++ b/internal/unifiedresources/host_disk_usage.go@@ -0,0 +1,21 @@+package unifiedresources++import "github.com/rcourtman/pulse-go-rewrite/internal/models"++// HostDiskUsagePercent uses the same first-filesystem admission as the+// canonical host metric. Missing or unavailable inventory is not measured zero.+func HostDiskUsagePercent(disks []models.Disk) (float64, bool) {+ if len(disks) == 0 || !hasObservedDiskUsage(disks[0]) {+ return 0, false+ }+ return disks[0].Usage, true+}++// DiskUsagePercent preserves canonical metric presence, including provider+// readings without agent filesystem inventory. Reporting freshness is separate.+func (v HostView) DiskUsagePercent() (float64, bool) {+ if v.r == nil || v.r.Metrics == nil || v.r.Metrics.Disk == nil {+ return 0, false+ }+ return v.r.Metrics.Disk.Percent, true+}diff --git a/internal/unifiedresources/metrics.go b/internal/unifiedresources/metrics.goindex e691009e02..737cb71ff8 100644--- a/internal/unifiedresources/metrics.go+++ b/internal/unifiedresources/metrics.go@@ -112,12 +112,9 @@ func buildHostMetricPayload(percent := percentFromReportedPercent(memory.Usage)metrics.Memory = &MetricValue{Used: &memory.Used, Total: &memory.Total, Percent: percent, Unit: "bytes", Source: source, Observation: memory.Observation}}- if len(disks) > 0 {+ if percent, ok := HostDiskUsagePercent(disks); ok {disk := disks[0]- if hasObservedDiskUsage(disk) {- percent := percentFromReportedPercent(disk.Usage)- metrics.Disk = &MetricValue{Used: &disk.Used, Total: &disk.Total, Percent: percent, Unit: "bytes", Source: source}- }+ metrics.Disk = &MetricValue{Used: &disk.Used, Total: &disk.Total, Percent: percent, Unit: "bytes", Source: source}}setNetworkAndDiskIOMetricsHost(metrics, netInRate, netOutRate, diskReadRate, diskWriteRate, source)return metricsdiff --git a/internal/unifiedresources/metrics_test.go b/internal/unifiedresources/metrics_test.goindex cd487ebe8f..d035bbedfa 100644--- a/internal/unifiedresources/metrics_test.go+++ b/internal/unifiedresources/metrics_test.go@@ -663,3 +663,44 @@ func TestGuestMemoryClusterResourcesIsTypeAware(t *testing.T) {}}}++func TestHostDiskUsagePercentMatchesCanonicalPresence(t *testing.T) {+ for _, tc := range []struct {+ name string+ disks []models.Disk+ present bool+ value float64+ }{+ {name: "absent"},+ {name: "no capacity", disks: []models.Disk{{}}},+ {name: "unavailable", disks: []models.Disk{{Total: 100, Usage: -1}}},+ {name: "invalid", disks: []models.Disk{{Total: 100, Usage: math.NaN()}}},+ {name: "infinite", disks: []models.Disk{{Total: 100, Usage: math.Inf(1)}}},+ {name: "inconsistent bytes", disks: []models.Disk{{Total: 100, Used: 101, Usage: 0}}},+ {name: "zero", disks: []models.Disk{{Total: 100, Free: 100}}, present: true},+ {name: "small percent", disks: []models.Disk{{Total: 1000, Used: 2, Usage: 0.2}}, present: true, value: 0.2},+ {name: "first filesystem only", disks: []models.Disk{{}, {Total: 100, Usage: 25}}},+ } {+ t.Run(tc.name, func(t *testing.T) {+ value, present := HostDiskUsagePercent(tc.disks)+ if present != tc.present || value != tc.value {+ t.Fatalf("disk usage = %v/%v, want %v/%v", value, present, tc.value, tc.present)+ }+ resource, _ := resourceFromHost(models.Host{ID: "disk-host", Disks: tc.disks})+ viewValue, viewPresent := NewHostView(&resource).DiskUsagePercent()+ if viewPresent != present || viewValue != value {+ t.Fatalf("canonical view = %v/%v, report = %v/%v", viewValue, viewPresent, value, present)+ }+ })+ }+ for _, resource := range []*Resource{nil, {}, {Metrics: &ResourceMetrics{}}} {+ if _, present := NewHostView(resource).DiskUsagePercent(); present {+ t.Fatal("missing canonical metric became zero")+ }+ }+ // Non-agent providers can report an aggregate without an agent inventory.+ provider := &Resource{Metrics: &ResourceMetrics{Disk: &MetricValue{Percent: 0}}}+ if value, present := NewHostView(provider).DiskUsagePercent(); !present || value != 0 {+ t.Fatal("provider's explicit zero requires no agent inventory")+ }+}