in new 4.99 builds, KMP is now deployed with readinessProbe and livenessProbe in the manager container:

      readinessProbe:
        httpGet:
          path: /readyz
          port: webhook-server
          scheme: HTTPS
          httpHeaders:
            - name: Content-Type
              value: application/json
        initialDelaySeconds: 10
        timeoutSeconds: 1
        periodSeconds: 10
        successThreshold: 1
        failureThreshold: 3
      name: manager
      command:
        - /manager
      livenessProbe:
        httpGet:
          path: /healthz
          port: webhook-server
          scheme: HTTPS
          httpHeaders:
            - name: Content-Type
              value: application/json
        initialDelaySeconds: 15
        timeoutSeconds: 1
        periodSeconds: 20
        successThreshold: 1
        failureThreshold: 3

It turns out, that on "real" clusters, these timers are too short and KMP doesn't have time to reach the ready state, and as a result the liveness probe causes the KMP pod to restart prematurely, casing a CrashLoopBackOff

100%, on clusters with real workloads (not CI/test clusters)

1. Install or upgrade CNV to the specified version
2. check the status of the KMP pod
3.

KMP pod is being crashlooped

KMP pod should be ready

caused by:
https://github.com/k8snetworkplumbingwg/kubemacpool/pull/540

This issue causes KMP not to function and blocks CNV upgrade.