While investigating issue [1] we've noticed a few problems with CNO error reporting on the ClusterOperator status [2]:

that's fine, but I think there are a couple bugs to write up:
1. when a panic happens, the operator doesnt' go degraded. This can definitely be done
2. when status cannot be updated, the operator should go degraded
3. when service network and/or clusternetwork in status is missing, the operator should go Available=false.

[1] https://github.com/openshift/cluster-network-operator/pull/1669
[2] https://coreos.slack.com/archives/CB48XQ4KZ/p1671207248527519?thread_ts=1671197854.825529&cid=CB48XQ4KZ

 4.13 and previous.

 Always

1. Cause a deliberate panic e.g. in the bootstrap code.

 Operator keeps getting restarted and is not Degraded.

 Operator goes Degraded.