Encountered an issue where a customer manually stopped machines for a cluster deployment with an unreachable condition and the powerState within cd status gets stuck in a WaitingForMachines state.

Steps to reproduce:

1. Make cluster's kubeconfig invalid in some way to have an unreachable condition.
2. Stop a machine for the cluster with no explicit cd.spec.powerState set, eg. hibernation controller expects to start the machines it found stopped.
3. Observe the hibernation controller start the machine and the cd powerState being set to WaitingForMachines. This condition will never clear despite the machines having successfully started.

  - lastProbeTime: "2023-03-13T20:26:43Z"
    lastTransitionTime: "2023-03-13T20:26:43Z"
    message: 'Waiting for cluster machines to start. Some machines are not yet running:
      i-0142d17411ffb0202 (step 1/4)'
    reason: WaitingForMachines
    status: "False"
    type: Ready 

https://redhat-internal.slack.com/archives/CE3ETN3J8/p1678735174858429