[Risk Mitigation] Create backup for ClusterRoleBinding, OIDCConfiguration and ShootSpecs on all SKR clusters on PROD before running migration script + implement restore script

[tobiscr]

Description

In case of failures, we have to be able to restore the previous cluster state. KIM is modifiying two resources and we have to create backups of them:

• Shoot-Spec on Seed-Cluster
• ClusterRoleBindings on SKR
• OIDCConfiguration on SKRs

In an emergency case, we have to be able to apply the backup as fast as possible. Therefore, a restore script for the resources located on an SKR required (ClusterRoleBinding and OIDCConfiguration).

AC:

• Before the migration from Provisioner to KIM is started, a backup is created which dumps
  • Shoot-Specs from Gardener Seed Cluster
  • ClusterRoleBinding resources from SKRs
  • OIDCConfiguration resources from SKRs
• Restore script is in place quich recovers the backed resouce (ClusterRoleBinding and OIDCConfiguration) on an SKR cluster

Reasons

Mandatory for restoring a defective SKR cluster.

Attachments

[akgalwas]

Workplan:

• Refactor code to introduce backup and restore tools
• Handle shoots
  • Backup
  • Restore
• Handle CRBs and OIDC
  • Backup
  • Restore
• Move the labeling CRB as deprecated from the migrator to the backup tool
• Prepare scripts for migration process
  • backup and switching runtime to be controlled by KIM
  • restoring from backup
• Prepare READMEs