ropee deployment on kubernetes

[benoittoulme]

Hi, thank you for your work on ropee!

We have ropee running in a kubernetes cluster, I do not know if you tried to do that before. kubernetes requires a readiness / liveness probe, which needs to return 200 (or ropee will be killed by k8s)

We tried to use /metrics as probe, but that caused some issues, maybe because prometheus is using the same endpoint to collect metrics. So we implemented a dumb /health that... returns 200.

Have you ever tried ropee on k8s? Do you expect /metrics to work as liveness probe call? Do you have ideas about better healthchecks than just returning 200?

I can help with an upstream PR if you are interested.

[kebe7jun]

@benoittoulme Hi, I have ever deployed ropee on Kubernetes, but no liveness probe added. It is a good idea to add a /healthz endpoint for liveness or readyness probe. Looking forward for your PR. ; )