Restore to re-configure workers

The main k0s retore is a full restore of the control plane which assumes that the control plane address (the address where workers connect to) does not change. If the address changes, everything on the controlplane is able to reconcile the state (kube api endpoints etc.) but the workers insist of connecting back to the same old address they used prior to crash-and-restore.

k0sctl could loop through all the workers and "fix" the connection address if needed. It should be mainly looking at the kubelet.conf (the kubeconfig for kubelet) and change the server address for the kube api server.

Start by tracing the main k0s restore flow in k0sctl and inspect each worker's kubelet.conf handling. Reproduce a restore where the control-plane address changes, then verify that workers use the new kube API server address afterward.

Written by the indexing model from the issue text.