Detect conditions when workers need to be restarted or re-joined

As mentioned by @twz123 in https://github.com/k0sproject/k0sctl/issues/721#issuecomment-2137655712 - there are situations that could be consistently detected by k0sctl that require that the existing workers are restarted / reconfigured / rejoined.

This is something k0sctl should handle gracefully.

Start by reviewing issue #721 and the referenced comment for the conditions that should trigger worker handling. Then trace how k0sctl currently manages existing workers and identify where restart, reconfiguration, or rejoin decisions belong. Done means consistently detected conditions are handled gracefully without requiring manual intervention.

Written by the indexing model from the issue text.