Issue: ovn-kubernetes-master pods are in crash-loop back-off state and forcing new leader elections every few seconds/minutes.
cluster is degraded, operators impacted.

Kube-apiserver is reporting a requeuing error indcating it's likely flooded

openshift-multus has a scheduler pod down, that cannot bind an IP with error: "error adding container to network ovn-kubernetes cni request failed with status 400" --> https://access.redhat.com/solutions/6646561 [restarting pods/databases, does not mitigate issue]

4.12.21

customer-specific

1.
2.
3.

cluster upgrade should not stall/crash-loop ovnkube-master pods

next update (internal) will include logs/sample data + troubleshooting taken so far. Impact is high for customer, production down.