Planning Grafana Mimir capacity

Grafana
,
1

Hello,

At our aks cluster we plan Mimir capacity which we run in microservice mode.

We want to estimate the resource utilization and followed the site: Grafana documentation
Problem is that for monitoring we use grafana/k8s-monitoring-helm chart and there is no any “prometheus_tsdb_head_series” metric for check current active series.

Does exist different way or metric how to estimate our usage?

Thanks.

2

Hi,

There’s a brute way count({__name__=~".+"}) but if you have lots of series, it can quickly kill your instance, so do that when nothing else works :smile:. In the comment to this stack overflow answer they say, there might be prometheus_agent_active_series metric - do you have it?

Edit: another answer says that there’s a status table in prometheus UI - maybe you’ll be able to get there?

3

Hello,
the problem is that all our metrics are scraped via Allloy (grafana/k8s-monitoring-helm) into Mimir instances (Distributro, Ingester,…). Only what we have are cortex metrics from Service Monitor. Is there a way to estimate the usage by them?