This message was deleted.

Can you check the logs of clickhouse ?

There are no obvious errors in the clickhouse logs.

From what version did you upgrade ?

where should i look at clickhouse metrics

Upgraded from 0.36 to 0.37

We don’t have clickhouse changes in 0.37, Can you check if you have changed the retention period recently https://signoz.io/docs/userguide/retention-period/

I modified the default settings, metics are retained for 15 days, traces are retained for 15 days, and I did not collect logs.

Did you do it recently ?

Yes, after the upgrade, I cleared the clickhouse data and then modified the configuration of signoz

By clearing the data do you mean deleting entire clickhouse data or something else ?

No, that PR wouldn't be the reason. If it were to be the reason everyone else would also see the issue (including our customers), I wouldn't expect systems to cause a spike. Any time you change something with TTL then it's going to cause a lot of CPU resource consumption.

I decided to lower the background_pool_size, MergeMutate threads became fewer, and the overall CPU usage decreased, but I don’t know what other effects it had.

Don't change TTL often on large tables.

@nitya-signoz Yes, I cleared all the data of signoz’s clickhouse instance.

You said you modified the ttl recently here https://signoz-community.slack.com/archives/C01HWQ1R0BC/p1706089767556039?thread_ts=1706066646.463219&cid=C01HWQ1R0BC.

@Srikanth Chekuri I mean, I suspect signoz 0.37 or signoz-schema-migrator:0.88.8 increased the TTL of some tables.

No, there was no such change