Open hjsauce opened 1 week ago
Also seeing this issue and seems related to this issue: https://github.com/argoproj/argo-cd/issues/19854
Facing the same issue. We have an alert to monitor argocd_cluster_connection_status metric to detect this situation , however the metrics also stops reporting for the affected cluster when this occurs. Hence alert is not firing. We are using consistent-hashing for sharding.
Checklist:
argocd version
.Describe the bug
After an update to 2.12.3, I noticed that some of the clusters were in an unknown state. Note that this does not happen immediately. It happens over time.
I am using ArgoCD in HA mode with
consistent-hashing
. I have configured 3 replicas for the controller.I am currently able to fix the issue by restarting the application-controller statefulset.
To Reproduce
Expected behavior
ArgoCD works as expected and I can sync applications
Screenshots Cluster errors Application errors
After restart of the controller statefulset
Version
Logs