My customer reported me the following feedback : 

We have a dev environment that was not properly configured, and RHOSO operator updated by itself. With the updated operator came this issue: 

barbican - operator created another stateful set and did not remove the older one - this caused intermittent issues when request came to different backends. Im unable to find the screenshot, but from what I recall the new set had names like barbican-worker instead of barbican-worker-worker which we have now.

It is a little disturbing that operator update alone can make breaking changes and that this is not documented - I can only assume that barbican change is related to https://issues.redhat.com/browse/RHOSSTRAT-526. The dev environments are running 18.0.6-20250317.1 and 18.0.8-20250505.2. The operator was rolled back to the previous working version - 1.0.11. 1.0.12 come with 2 mentioned changes.

--> Dev environments are now broken and lot of doubts about updating the production.