When using MachineHealthCheck and ControlPlaneMachineSet together to verify automatic recovery. for example, one of the ControlPlanes is stopped and automatically restored.

The Node remains in NotReady status and the Machine remains in Deleting status. In my environment I am verifying this in an AWS IPI environment, if I manually delete the instance using AWSconsole, the Node and Machine are deleted afterwards. Similar events have been reported in the vSphereIPI environment.

OCP4.13

1.Setting up ControlPlaneMachineSet and MachineHealthCheck 
2.Making one ControlPlane to NotReady (Ex: Using # systemctl stop kubelet)
3.Deleting the instance (or VM) manually.

After step 2, the machine remains in deleting status and the node also remains in Notready status. They will disappear by performing step 3.

Is this expected behavior, and is it being intended to be removed by the administrator? Or is there some kind of bug that is stopping the process?

It makes sense that since ControlPlane also stores etcd's DB, instances and VMs are kept due to conservative design. However, if that is the case, I would like the documentation to clearly state that manual deletion by the administrator is required.On the other hand, it was reported that VMs are automatically deleted when using machineset and machinehealthcheck on workers and others. Therefore, ControlPlaneMachineSet could be considered an unintended behavior.