First of all, i’m new here, please let me know if this is not the correct place to discuss Grafana Mimir. (Pointers to the correct place would be awsome in that case)
In my company, we’re currently running a Grafana, Prometheus setup with many teams using many Prometheus instances.
We want to migrate to Grafana Mimir (self hosted in Kubernetes)
When using High-availability deduplication, i hit some snags and was trying to see which of my Prometheus Instances was the currently active one. To avoid the deployment of a whole separate token store provider just for ha tracking, we’re using the inmemory provider.
-
What Mimir version and what operating system are you using?
Mimir 2.8.0 on Kubernetes -
What are you trying to achieve?
I’m trying to get visibility on which of the Prometheus High-availability Instances I’m accepting metrics from -
How are you trying to achieve it?
It would be interesting to be able to query a metric that tells me by some measure which Prometheus replica is currently the active one. But i can not find a way to do this with high availability using inmemory provider.
It obviously works using an external, ha_tracker like consul. But even then, one has to go to that store to see which replica the current metrics are coming from.