v.25.12Improvement

Avoid overwhelming Prometheus

Official CH changelog text
We want to run keepers in a separate server processes and, to avoid overwhelming Prometheus in large regions, we should continue to expose only keeper-related metrics. #90244 (Miсhael Stetsyuk).
Introduce the ability to run keepers as separate server processes with continued exposure of only keeper-related metrics to avoid overwhelming Prometheus in large regions.

Why it matters

This feature addresses the problem of running keepers in isolated server processes while managing Prometheus monitoring efficiently. By limiting exposed metrics to keeper-related data only, it prevents Prometheus from being overwhelmed with high volumes of metrics in large deployments, improving monitoring scalability and clarity.

How to use it

Deploy keepers as separate server processes in your ClickHouse cluster. The server will automatically expose only keeper-related metrics to Prometheus, ensuring efficient monitoring without additional configuration required for metric filtering.

Related resources

PreviousNext