Deleting everything now fails with

dial tcp: lookup <http://xxx.gr7.us-east-1.eks.amazonaws.com|xxx.gr7.us-east-1.eks.amazonaws.com>: no such host

So - having to replace the cluster if you make significant network topology changes is expected.

gets stuck since it cant delete the resources for the now deleted cluster

This on the other hand isnt.

Deleting everything now fails with

dial tcp: lookup <http://xxx.gr7.us-east-1.eks.amazonaws.com|xxx.gr7.us-east-1.eks.amazonaws.com>: no such host

Nor is this.

Ah yes, that looks like a likely root cause

For this in particuar:

Deleting everything now fails with

dial tcp: lookup <http://xxx.gr7.us-east-1.eks.amazonaws.com|xxx.gr7.us-east-1.eks.amazonaws.com>: no such host

It sounds like you have managed to delete your cluster without deleting the Kubernetes resources installed in it. That should not be possible, but I believe due to 2948 above it could be if you had otehr failures previously.

That sounds about right

In this case - the easiest remedy for now is to

pulumi stack export > stack.json

and then remove all of the Kubernetes resources that no longer really exist, then

pulumi stack import < stack.json

.

Ah, the second one sounds a bit nicer