-
Bug
-
Resolution: Duplicate
-
Major
-
4.19.0
-
None
TRT-2049 found that the method for collecting watch counts was changed to track only deprecated api calls. This left a gap in monitoring until it was discovered and the logic moved to audit log analysis monitor tests.
[sig-arch][Late] operators should not create watch channels very often [apigroup:apiserver.openshift.io] [Suite:openshift/conformance/parallel]
is showing regressed for vshpere and gcp serial jobs currently.
Now that we can collect the data again we need to analyze the current base line, backport the fix to 4.18 and potentially adjust the caps.
- duplicates
-
OCPBUGS-55897 operators should not create watch channels very often regression
-
- MODIFIED
-
- is cloned by
-
OCPBUGS-55785 operators should not create watch channels very often regression
-
- MODIFIED
-
- relates to
-
OCPBUGS-54928 CI: [sig-arch][Late] operators should not create watch channels very often: {Operator "ingress-operator" produces more watch requests than expected: watchrequestcount=841, upperbound=840, ratio=1 }
-
- Closed
-
- links to