Hive should re-attempt in case of GeneralOperatorDegraded, so that the clusters don't run into an error state (bad customer experience). 

For example, if OCPBUGS-17062 occurs, we can repair the cluster by replacing a worker node.
The OCM state, however, cannot be bounced out of error, thus we currently have to send customers a generic SL to retry: "Your install failed due to an intermittent issue. Please retry installation." 

Note: Ideally, we should be able to re-initiate installs (kick them off again where they stopped) even if a cluster is already in error state. I wonder if that's feasible? 

Done:

• Have Hive attempt a re-install for GeneralOperatorDegraded Failures while we pursue the issue upstream