When a user supplies a invalid CR or updates a existing broker deployment with a invalid/incomplete patch the broker pod will be in a CrashLoopBackOff state and it blocks the operator it handle further patches that may fix the broker.

It would be better if there is a way to avoid this.