Merge pull request 'fix: [alert] OpenBaoQuorumLost (openbao) (#920)' (#921) from alert-fix/issue-920 into master

Reviewed-on: k.renner/homelab#921
2026-09-23 15:59:33 +00:00
parent 80816d7433 87083c46e7
commit 6bb8ef4776
1 changed files with 7 additions and 3 deletions
@@ -57,10 +57,14 @@ spec:
          # the ClusterSecretStore goes InvalidProviderConfig and no
          # ExternalSecret can refresh.
          #
-          # No window: below quorum the secret store is down and every
-          # ExternalSecret stalls with it, so waiting only delays the page.
+          # Short window: readiness is only a proxy for quorum. A one-second
+          # node NotReady flap makes the node lifecycle controller mark that
+          # node's pods not-ready for ~30s while raft keeps its leader and
+          # serves (issue #920, during a rolling reboot with one replica
+          # already evicted). A really sealed or crashed majority persists,
+          # and existing Secrets keep their last values, so 2m costs little.
          expr: max by (namespace, statefulset) (kube_statefulset_status_replicas_ready{namespace="openbao", statefulset="openbao"}) < 2
-          for: 0m
+          for: 2m
          labels:
            severity: critical
          annotations: