#support

Self-hosted Signoz Retains All Traces Beyond Limit

TLDR Prashant has a problem with Signoz retaining traces beyond the default limit, causing storage to be full. Srikanth suggests using "OPTIMIZE TABLE ..." command in Clickhouse to trigger cleanup.

Powered by Struct AI
6
2mo
Solved
Join the chat
Jul 23, 2023 (2 months ago)
Prashant
Photo of md5-9666afc0ebaac245df53cf43559d3630
Prashant
01:49 PM
Hey in my self hosted signoz setup. Signoz is retaining all traces it should clear up metrics that are older than a than a week, which is the default behavior. This is leading to my storage being full. I installed signoz using the default chart on my kubernetes cluster.
Jul 24, 2023 (2 months ago)
Srikanth
Photo of md5-ce04a9988e2fd758a659dc55be6f2543
Srikanth
04:19 AM
That shouldn’t be the case, How long did it retain beyond the TTL limit?
Prashant
Photo of md5-9666afc0ebaac245df53cf43559d3630
Prashant
06:02 AM
The trace time limit is 7 days. I can see traces from 27th of June. Is there a way to force a cleanup?
Srikanth
Photo of md5-ce04a9988e2fd758a659dc55be6f2543
Srikanth
06:47 AM
Can you share one such example trace and confirm if you haven’t updated the retention from settings?
Prashant
Photo of md5-9666afc0ebaac245df53cf43559d3630
Prashant
07:26 AM
Here’s the trace details screenshot and the settings for traces.
Image 1 for Here’s the trace details screenshot and the settings for traces.Image 2 for Here’s the trace details screenshot and the settings for traces.
Jul 25, 2023 (2 months ago)
Srikanth
Photo of md5-ce04a9988e2fd758a659dc55be6f2543
Srikanth
02:47 AM
That’s unusual. You need to exec into clickhouse and run OPTIMIZE TABLE ... FINAL to force trigger the cleanup. Run this on the signoz_index_v2 table.