Cleanup Grace Period

Deleting a SveltosCluster normally removes it right away: unlike a ClusterAPI Cluster, there is no infrastructure provider holding a finalizer on it, so nothing keeps the object around during teardown.

That is fine for the SveltosCluster object itself, but it also means nothing else gets a chance to react while it still exists. In particular, the event-driven cleanup pattern that lets you run a Job in the management cluster while a cluster is deleting depends on the cluster object still being there, with a deletionTimestamp, long enough to be evaluated.

SveltosCluster.Spec.CleanupGracePeriod closes that gap. When set, Sveltos holds its own finalizer on the SveltosCluster for that long after deletion is requested, before actually removing it:

apiVersion: lib.projectsveltos.io/v1beta1
kind: SveltosCluster
metadata:
  name: prod-cluster
  namespace: civo
spec:
  cleanupGracePeriod: 10m

With this set, deleting prod-cluster keeps the SveltosCluster present (with metadata.deletionTimestamp set) for 10 minutes before Sveltos removes it, instead of disappearing immediately.

By itself, this only delays removal; it does not run any cleanup. Pair it with an EventSource/EventTrigger using includeDeletingResources: true against this SveltosCluster to actually act during that window. See Event Driven Cleanup on Cluster Deletion.