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
Integrate reviewed metric observation continuity repair
pulse-triage[bot]eff676c746c0
diff --cc internal/alerts/node.goindex 861e615ee8,978fa9aad2..6698bdf64c--- a/internal/alerts/node.go+++ b/internal/alerts/node.go@@@ -93,37 -93,27 +93,37 @@@ func (m *Manager) CheckNode(node modelsm.clearNodeOfflineAlert(node)// Check each metric (only if node is online and reachable)- // Check for host agent deduplication: if a host agent is linked to this node,- // CPU, memory and disk usage alerts belong to the agent resource (CheckHost- // evaluates them), so the node releases its own copies instead of alerting twice.- if m.hasHostAgentForNode(node.ID) {- m.releaseNodeMetricAlerts(node, "cpu", "memory", "disk")+ // Check for host agent deduplication: a usage metric that a linked host+ // agent evaluates belongs to the agent resource (CheckHost raises it), so+ // the node releases its own copy instead of alerting twice. A metric the+ // agent does not evaluate (agent alerts disabled, threshold off, usage+ // unknown) stays with the node, so the machine is never left unmonitored.+ link, _ := m.linkedHostAgentForNode(node.ID)+ input := &UnifiedResourceInput{+ ID: node.ID,+ Type: "node",+ Name: node.Name,+ Node: node.Name,+ Instance: node.Instance,+ }+ released := make([]string, 0, 3)+ if link.cpu {+ released = append(released, "cpu")} else {- var memoryMetric *UnifiedResourceMetric- if node.Memory.HasKnownUsage() {- memoryMetric = &UnifiedResourceMetric{Percent: node.Memory.Usage}- }- m.evaluateUnifiedMetrics(&UnifiedResourceInput{- ID: node.ID,- Type: "node",- Name: node.Name,- Node: node.Name,- Instance: node.Instance,- CPU: &UnifiedResourceMetric{Percent: node.CPU * 100},- Memory: memoryMetric,- Disk: &UnifiedResourceMetric{Percent: node.Disk.Usage},- }, thresholds, nil, "temperature") // checkNodeTemperature owns its observation gaps.+ input.CPU = &UnifiedResourceMetric{Percent: node.CPU * 100}+ }+ if link.memory {+ released = append(released, "memory")+ } else if node.Memory.HasKnownUsage() {+ input.Memory = &UnifiedResourceMetric{Percent: node.Memory.Usage}}+ if link.disk {+ released = append(released, "disk")+ } else {+ input.Disk = &UnifiedResourceMetric{Percent: node.Disk.Usage}+ }+ m.releaseNodeMetricAlerts(node, released...)- m.evaluateUnifiedMetrics(input, thresholds, nil)++ m.evaluateUnifiedMetrics(input, thresholds, nil, "temperature") // checkNodeTemperature owns its observation gaps.// CPU temperature stays with the node even when a host agent runs on it:// CheckHost has no CPU temperature metric, and the node poll already