test/e2e: Cluster deprovisioning not working

[stevenhorsman]

As discovered when testing the docker-provider e2e, but I suspect it's not unique to that provider. When I didn't specify TEST_TEARDOWN=no, the cluster is supposed to be removed at the end of the test, however that didn't happen due to the peerpod-ctrl uninstall failing.

Normally it looks like:

time="2024-07-01T11:40:06-07:00" level=info msg="Delete the peerpod-ctrl deployment"
ok  	github.com/confidential-containers/cloud-api-adaptor/src/cloud-api-adaptor/test/e2e	367.040s

and silently fails, but with tracing on we can see that the peerpod-ctrl errored:

Error from server (NotFound): error when deleting "STDIN": namespaces "confidential-containers-system" not found
Error from server (NotFound): error when deleting "STDIN": serviceaccounts "peerpod-ctrl-controller-manager" not found
Error from server (NotFound): error when deleting "STDIN": roles.rbac.authorization.k8s.io "peerpod-ctrl-leader-election-role" not found
Error from server (NotFound): error when deleting "STDIN": rolebindings.rbac.authorization.k8s.io "peerpod-ctrl-leader-election-rolebinding" not found
Error from server (NotFound): error when deleting "STDIN": services "peerpod-ctrl-controller-manager-metrics-service" not found
Error from server (NotFound): error when deleting "STDIN": deployments.apps "peerpod-ctrl-controller-manager" not found
make[1]: *** [Makefile:182: undeploy] Error 1
make[1]

There are probably a few bugs here:
1. That the `make -C ../peerpod-ctrl/ undeploy` is failing
2. That the e2e test silently fails during teardown
3. That an error in `Delete the peerpod-ctrl deployment` causes the teardown to exit and the cluster to not be created