Add cluster_control_plane_unhealthy inhibition. (#1155)

Co-authored-by: Quentin Bisson <[email protected]>
giantswarm · May 6, 2024 · 1e34b67 · 1e34b67
1 parent 68b0304
commit 1e34b67
Show file tree

Hide file tree

Showing 2 changed files with 13 additions and 0 deletions.
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -17,6 +17,10 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 - Fix etcd alerts for Mimir.
 - Add missing labels for apiserver alerts.
 
+### Added
+
+- Add `cluster_control_plane_unhealthy` inhibition.
+
 ## [3.13.1] - 2024-04-30
 
 ### Removed

diff --git a/helm/prometheus-rules/templates/alerting-rules/inhibit.management-cluster.rules.yml b/helm/prometheus-rules/templates/alerting-rules/inhibit.management-cluster.rules.yml
@@ -95,4 +95,13 @@ spec:
         instance_state_not_running: "true"
         team: phoenix
         topic: status
+    - alert: InhibitionControlPlaneUnhealthy
+      annotations:
+        description: '{{`Control plane of cluster {{ $labels.cluster_id }} is not healthy.`}}'
+      expr: capi_kubeadmcontrolplane_status_condition{type="ControlPlaneComponentsHealthy", status="False"} == 1 or capi_kubeadmcontrolplane_status_condition{type="EtcdClusterHealthy", status="False"} == 1 or capi_kubeadmcontrolplane_status_condition{type="Available", status="False"} == 1
+      labels:
+        area: kaas
+        cluster_control_plane_unhealthy: "true"
+        team: phoenix
+        topic: status
 {{- end }}