KIM should set annotations to reduce the grace-periods when deleting a cluster

[tobiscr]

Description

We were already investigating the force-deletion annotation of Gardener (see #103 ) which turned out to be not a perfect match for our purposes.

Alterantively, we could reduce the grace-period timespans to minimise the waiting time. See https://pages.github.tools.sap/kubernetes/gardener/docs/landscapes/live/gardener/shoot_cleanup/ for getting an overview of available annotations.

AC:

• Run a POC which is using these annotations and set them to <5min. Goal is to identify a value which is high enough to delete a normal cluster (which is in a halthy state) without a risk of generating orphan infrastructure resources, but small enough to avoid unnecessary waiting times if the deletion hangs.
• Extend KIm to set these Gardener annotaitons to reduce graceful periods to a meaningful value

Reasons

Reduce waiting time if a cluster hangs during the deletion.

Attachments