Use persistent volume for metrics

[jochen-schuettler]

A disclaimer: Please understand I'm a novice in these matters and may have a shaky grasp on some concepts.
Whenever Kepler's pod is cleared and reopened, the metrics are lost. I would like Kepler to use a persistent volume for Prometheus metrics. I guess this is a mixed problem between opening such a volume in K8s and using it in Kepler. So the feature would be to add a parameter to activate and refer to the persistent volume. And the standard helm chart given to install Kepler should introduce such a volume and parametrize Kepler accordingly. That means it can still be run without persistent volume, but usually it would be installed with persistent memory and use it.

[jochen-schuettler]

I found out how to modify my helm chart deploying grafana, namely by adding:

  persistence:
    type: statefulset
    enabled: true

But as this is out of scope of the kepler project, this discussion never made any sense here. So I'm closing this.