Readiness probes of some components not correct

[TilBlechschmidt]

🐛 Bug description

It appears that some components do not propagate the correct readiness probe state regarding connectivity to the Redis server. It seems to be affecting the manager, orchestrator, and gangway. The collector and api probably suffer the same issue but in the observed scenario they kept on crashing because the mongodb server was unavailable.

🦶 Reproduction steps

Steps to reproduce the behavior:

1. Deploy a webgrid fresh
2. Make sure the redis and/or MongoDB don't come up
3. Watch it burn 🔥

🎯 Expected behaviour

This is more of a philosophical discussion on whether the software should crash upon encountering an error or just report a negative readiness state. Probably the latter, however, even that is currently not given. Redis connectivity should be reflected in the readiness state!

📺 Screenshots

[TilBlechschmidt]

After giving it a few minutes to settle down, it appears as though the api and collector crash (as expected), the gangway correctly report the readiness, and the manager and orchestrator behave incorrectly.