You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I've generally been chasing down Application controller perf issues that seem to be rooted in extremely active clusters cause excessive reconciliation triggering. A few of the items that seem to trigger it are Endpoints, EndpointSlices, DaemonSet updates largely driven by cluster autoscaling.
To Reproduce
Deploy a service with endpoints, check UI for endpoints object
Expected behavior
Services should not be tracking the endpoint object.
It appears adding endpoints explicitly to resource.exclusions has them ignored by ArgoCD.
Added to 3.0 milestone with #20013 to change the default configurable settings. The current behavior is expected not to exclude Endpoints by default, but doc was not up-to-date.
Checklist:
argocd version
.Describe the bug
The resource exclusion docs seem to indicate that
Endpoints
should be excluded/not presented and tracked in ArgoCD: https://argo-cd.readthedocs.io/en/stable/operator-manual/declarative-setup/#resource-exclusioninclusionI am still seeing them in our ArgoCD UI however.
I've generally been chasing down Application controller perf issues that seem to be rooted in extremely active clusters cause excessive reconciliation triggering. A few of the items that seem to trigger it are Endpoints, EndpointSlices, DaemonSet updates largely driven by cluster autoscaling.
To Reproduce
Deploy a service with endpoints, check UI for endpoints object
Expected behavior
Services should not be tracking the endpoint object.
It appears adding endpoints explicitly to
resource.exclusions
has them ignored by ArgoCD.Screenshots
Version
References:
argoproj/gitops-engine#469
#8175
The text was updated successfully, but these errors were encountered: