AWS WordPress example: resources cannot be cleaned up, because of orphan resources about provider-aws

[negz]

My guess is the issue here is to do with the relationship between the KubernetesApplication and the KubernetesCluster it was scheduled to.

My recollection is that if the KubernetesApplicationResource controller is trying to delete a KubernetesApplicationResource but finds it can't connect to its scheduled KubernetesCluster because it no longer exists it assumes there's nothing that can be done, and considers the KubernetesApplication deleted. So I'm guessing what's happening here is:

1. Deploying the KubernetesApplication creates an ELB in the VPC used by the EKSCluster bound to the scheduled KubernetesCluster.
2. When deleting everything all at once the KubernetesCluster happens to get deleted before the KubernetesApplication, so it has no way of deleting the remote resources created by its KubernetesApplicationResources - they keep running. The EKSCluster itself is deleted, but nothing cleans up its services (and thus ELBs).
3. We now can't delete the VPC because there are orphaned ELBs in it.

[muvaf]

This is irrelevant since these CRDs are removed now.