cortex
cortex copied to clipboard
Document blocks storage per-tenant retention
The PR https://github.com/cortexproject/cortex/pull/3879 introduced per-tenant retention for the blocks storage but we currently don't have any documentation about it. We should mention it in the doc.
Ideas:
- We could mention it in
docs/blocks-storage/compactor.template
- We could have a "Retention" guide under
docs/blocks-storage
mentioning it + mention that a bucket policy along with-querier.max-query-lookback=<retention period + small delta>
may be another alternative but wouldn't provide a per-tenant granularity
Hey @pracucci, I am currently setting up the retention policy for my cortex, and I am interested in contributing to this issue. I am think that we could do the following:
- Add the global retention config in
docs/blocks-storage
with-querier.max-query-lookback=<retention period + small delta>
. - Add per tenant config in runtime-config, add a section
compactor & block storage
like what we have for Ingester, Distributor & Querier limits. - Link 2 from 1, so that user can navigate to per-tenant config if they are interested in that.
WDYT?