Min RK

Since you can run repo2docker in docker by mounting the docker socket (no mount necessary with docker-machine), this should be quite doable: ```bash docker run --rm -it -v /var/run/docker.sock:/var/run/docker.sock -e...

We should also probably be polling `/api/status` to check that the Hub is up, and showing a restart link if we start getting errors.

I think a great enhancement would be when a stopped container is about to resume, go through and check for fields that differ from what we would be starting and...

> KubeSpawner should stop using gen.maybe_future and instead rely on from jupyterhub.utils import maybe_future Yes! That's the right thing to do. If you want, though, I don't think kubespawner needs...

Makes sense, I think a general ```python async def retrying_api_request(): try: return (await real_api_request()) or True # ensure truthy, or is the response always truthy? except ApiException as e: if...

I think it might be. How would you handle that? A single app-wide token? Or per-route tokens? Checking cookies vs headers, etc.? It's a bit of a can of worms...

Or were you thinking of token-based routing (i.e. specifying the target based on token)? That's an interesting idea. I'm not sure quite how it would fit.

If you're already using another layer to terminate SSL, perhaps it makes sense to do the http->https redirect there instead of the proxied service? Still, #187 adds a `--redirect-to` option...

Since user pods should not generally be automatically restarted, I think Unknown should probably be treated as a dead pod and cleaned up the same as if it has exited...

But because this is likely an issue with the cluster, deleting a pod that is Unknown should probably come with some extra warning output about the state of the pod...