This repository was archived by the owner on Aug 4, 2021. It is now read-only.
This repository was archived by the owner on Aug 4, 2021. It is now read-only.
Disk space much larger than table sizes #48
Description
I'm running the pg_prometheus docker image (0.2.2-pg11) connected to Prometheus (2.15.2) via the prometheus-postgres-adapter (0.6.0). I've been running for a couple weeks like this and just looked at the disk usage. It's at 900GB, but the tables (metrics*) add up to only 1 GB.
I saw that autovacuum wasn't running, so I manually ran "vacuum full". It cleaned about 40 GB. I then set the autovacuum_vacuum_scale_factor to 0.01 to get it to run more often. It didn't run all weekend.
Anybody know why the disk usage is so much higher than the table size or why the vacuum stuff doesn't seem to be doing much of anything?