I would need to restart my Traefik instance deployed on Kubernetes using the community Helm chart. Though I use the Prometheus metrics exporter and I would rather not have all values reset. How can I restart my Traefik deployment without losing metrics data ? Where are they stored ?
Did anyone found how to achieve this? I'm wondering the same
I don't think that's possible. Traefik is stateless.
To be honest, the current way is the recommended way. The lightweight exporter just exposes metrics of the current instance of Traefik. Prometheus crawls the exporter and takes care of persistence. Prometheus is smart enough to deal with counters that reset to zero and summarizing metrics over multiple replicas of Traefik pods.
Why do you want to do this?