When a new configuration is picked up by the authentication operator, it rolls out a new revision of oauth-server pods. However, since the pods define `terminationGracePeriodSeconds`, the old-revision pods would still be running even after the oauth-server deployment reports that all pods have been updated to the newest revision, which triggers the authentication operator to report Progressing=false.

The above behavior might cause that tests (and possible any observers) that expect the newer revision would be still routed to the old pods, causing confusion.

4.13

Always

1. Trigger new oauth-server rollout
2. Observe the authentication operator reporting Progressing while also watching the number of pods in the openshift-authentication namespace

CAO reports Progressing=false even though there are more than the expected number of pods running.

CAO waits to report Progressing=false when only the new revision of pods is running in the openshift-authentication NS.

-