ci: use jupyterhub/action-k8s-await-workloads

[consideRatio]

While developing Helm charts, we need to await them to become ready before running additional test than just verifying it starts up properly. Doing this requires some logic that I've extracted into a GitHub action.

1. Ability to declare / inspect k8s on what resources to await (workloads, namespace)
2. Ability to opt in to failure on detected container restarts (max-restarts)
3. Ability to time out after a total amount of time (timeout)

More details in https://github.com/jupyterhub/action-k8s-await-workloads