\>I would expect if it was related to the amount of space left, postgresql would just crash.
It definitely can, I've seen it happen once when I ran out of space on my old VPS.
\>mess with pool size
Probably not necessary as I rarely see Postgres connection utilization go above 50% on this instance with the defaults. Could be an avenue to investigate if IO isn't the issue though.
\>timeout settings to see if it helps
Definitely would recommend if not done already. Here is my Pleroma config for that:
```
config :pleroma, Pleroma.Repo,
timeout: 30_000,
queue_target: 100,
queue_interval: 10000,
parameters: [
plan_cache_mode: "force_custom_plan"
]
```
\>You can run iostat to see what %iowait you're dealing with.
Another useful thing is enabling the "Detailed CPU time" setting in htop under display options. It then also shows iowait in the CPU utilization bar graphs.
Also the absolute classic of Pleroma DB issues is the default Postgres config that has not been [PgTuned](https://docs.pleroma.social/backend/configuration/postgresql/#pgtune). Or too much database bloat which can be seen from the Phoenix dashboard as mentioned. It's under Ecto Stats->Bloat.
