lower retention_period to a multiple of chunks.period (1x168h)
Had to make this change because of what I found in the documentation:
The retention period must be a multiple of the index / chunks table "period" (see period_config).
I also found out that our current settings would retain 14 days of data, even though we asked for 8, because of how the retention system works.
By lowering the chunks.period
to 24h
, I hope to be able to make sure OAS retains less data, although I'm not sure when it will start doing that.
In addition, I added a change that will allow me to override all the loki-settings with a configmap on the cluster so I can debug this problem without having to make a new MR for everything I want to try. This is the solution proposed in #873 (closed), but only applied to Loki for now, because my current problem is a great test case for it.