Change the asynchronous_metric_log check interval

Question

In the documentation can find the asynchronous_metrics_log default get metrics per minute, but in my clickhouse,there are a number of metrics got per second.

enter image description here

enter image description here

enter image description here

Query 1000 row only can take 3 second record.

SELECT * FROM system.asynchronous_metric_log order by event_time desc LIMIT 1000;

In grafana,Working 2 hour can got 4 millions record on the table.

enter image description here

I change the config about this but It's unuseful

enter image description here

How to reduce the count about asynchronous_metrics_log? I take the Prometheus to clickhouse,Is the problem about Prometheus?

Answer 1

There are a few configuration settings for this:

asynchronous_heavy_metrics_update_period_s default: 120 (seconds) https://clickhouse.com/docs/en/operations/server-configuration-parameters/settings#asynchronous_heavy_metrics_update_period_s

asynchronous_metrics_update_period_s default: 1 (seconds) https://clickhouse.com/docs/en/operations/server-configuration-parameters/settings#asynchronous_metrics_update_period_s

I'm not exactly sure how ClickHouse decided between what is a heavy metric or a (normal) metric, but assuming that decision was made based on resources required