Occasionally it looks like the nested checks fail when trying to test that the ingress controller is available. There is a --wait on the helm install of it, but it looks like it may not be waiting until it really is fully working.
Seems like everything is ready, but curl is failing sometimes. Like, kube-proxy isn't plumbing up the iptables rules quite fast enough. We probably need to put a retry loop around the curl.
Occasionally it looks like the nested checks fail when trying to test that the ingress controller is available. There is a --wait on the helm install of it, but it looks like it may not be waiting until it really is fully working.