Today, many CI jobs fail on installations. gather_extra tries to add detection of operator status to add more specific test details. For operators that do not have proper status, it adds specific junits named “operator conditions xxxx” for those operators. But oftentimes, this introduces more noise and turns a whole column of CR red. 

We had a discussion during the team sync and David proposed another go cmd in origin to do the analysis. 

• Start small with a golang tool or openshift-tests command to verify cluster health, i.e. do we have the right amount of workers and master nodes available?
• Progress to generating test results using more complicated logic
• When a set of operators fails that are tied to a specific cause, generate skipped junits for operators (with a reason), and a top-level test indicating the root cause
• Cases: all machines working, all nodes working, no unsigned CSR’s for masters/workers, etc.
• Operator tests have preconditions that skip if particular cases aren’t working