Skip to content

armgit / Pulse

Public Imported from GitHub
0

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.

Respect existing duplicate suppression in node handover control

pulse-triage[bot]da7b4707180b
diff --git a/internal/alerts/node_disk_observation_test.go b/internal/alerts/node_disk_observation_test.go
index 76f340b695..99f888dfb8 100644
--- a/internal/alerts/node_disk_observation_test.go
+++ b/internal/alerts/node_disk_observation_test.go
@@ -192,6 +192,8 @@ func TestNodeDiskUnknownLeavesIndependentMetricsAndAgentHandover(t *testing.T) {
m.config.NodeDefaults.CPU = &HysteresisThreshold{Trigger: 80, Clear: 70}
m.config.NodeDefaults.Memory = &HysteresisThreshold{Trigger: 80, Clear: 70}
m.config.MetricTimeThresholds = map[string]map[string]int{"all": {"disk": 0, "cpu": 0, "memory": 0}}
+ m.config.MinimumDelta = 2
+ m.config.SuppressionWindow = 5
m.mu.Unlock()
node := nodeDiskSample("site-pve", 95)
m.CheckNode(node)
@@ -217,13 +219,16 @@ func TestNodeDiskUnknownLeavesIndependentMetricsAndAgentHandover(t *testing.T) {
t.Fatal("handover lost agent disk incident")
}
m.HandleHostRemoved(host)
- // Handover release uses its real observation time. Later reports must
- // advance the synthetic clock too, rather than predate that release.
+ // These are later reports, not a replay of the original observation.
clock.Store(int64(time.Second))
m.CheckNode(gap)
if testHasActiveAlert(t, m, canonicalMetricStateID(node.ID, "disk")) {
t.Fatal("removing agent invented a node reading")
}
+ // The existing 2% minimum delta suppresses a new occurrence at the same
+ // 95% within five minutes. Use a genuinely changed measured value; this
+ // control tests restored node evidence, not bypassing duplicate policy.
+ node.Disk = nodeDiskSample(node.ID, 98).Disk
clock.Add(int64(time.Second))
m.CheckNode(node)
if !testHasActiveAlert(t, m, canonicalMetricStateID(node.ID, "disk")) {