we need to determine the best way to remove karpenter nodeclaims and the related nodes, as well as how we clean up when an HCP cluster is removed.

karpenter makes ownership relationships between nodes and nodeclaims, we should investigate this to determine which objects we can delete to create a cascade effect. eg can we remove nodepool and see all nodeclaims be removed.

related code about checking removal in the karpenter-operator https://github.com/openshift/hypershift/blob/main/karpenter-operator/controllers/karpenter/karpenter_controller.go#L127

we also need to check to see if we will need to confirm that instances are gone from the infrastructure.

note: double check the mechanics around deleting a nodepool, we want to ensure that nodepool is deleted first so that more nodeclaims are not generated.

DoD: document with options and a plan for the next cards