Changes for v2.5.8-patch1 

2.5.8 final RC

Update gke-operator to v1.0.1

[2.5] Bump wins version to v0.1.1

Bump gke-operator to RC14

Check cluster edit

Add cluster id param

Bump dashboard for v2.5.8-rc18

Add context to Ready condition while updating

[release/v2.5] Update tags for v2.5.8-rc17

[2.5] Rancher runtime image

There was some mixups with the RKE2 changes that caused the runtime
image to be set incorrectly.

Also, there were changes in RKE2 that caused problems with containerd
not being able to find the $HOME environment variable. That is also
fixed here.

[release/v2.5] Update tags for v2.5.8-rc16

[2.5.8] handle the system default registry in post-delete-hook-job better

[release/v2.5] Depend on linux-amd64 for drone docker img digests

[v2.5] Bump gke-operator to RC12

Fix nodepool name dereference

[release/v2.5] Pass drone image digest settings via env vars

When zonal GKE clusters are updating the control plane, the control
plane usually becomes unreachable. This is expected and normal
behavior[1]. However, when the HealthSyncer controller runs, it cannot
reach the API and sets ClusterConditionReady to false with an alarming
error message. This change does not eliminate or ignore the error but
adds more context to the error message to help users understand that
this is normal.

[1] https://cloud.google.com/kubernetes-engine/docs/concepts/cluster-upgrades#upgrading_manually

[2.5] Fix docker-runtime-publish

Don't crash if nodepool name is unset.