Memory Usage Metrics Issue

I am trying to monitor CockroachDB Memory Usage through Grafana Dashboard. There are a few documents I can find to refer like below:
https://www.cockroachlabs.com/docs/stable/monitoring-and-alerting.html#events-to-alert-on


But nothing mentioned which Prometheus Metric should I use for the Memory Usage Monitoring.

I saw someone’s example to divide container_memory_usage_bytes by container_spec_memory_limit_bytes. But this percentage result will always be near 100%. Why is this metrics calculation result so high? Is CockroachDB supposed to be using that high percentage memory all the time?

If the above calculation cannot reflect the real memory usage, what metrics and calculation should I use?

Thanks,
Derek

Hi @derek.z,

The metric you’re looking for is sys_rss. This should match with the memory usage graph on the hardware dashboard in the UI, and tracks the memory utilized.

It sounds like you’re looking to express this as a percentage of the total - is that right?