Pleroma performance issue #366

New Issue

Closed

opened 2022-11-03 10:24:16 +01:00 by muppeth · 1 comment

muppeth commented

2022-11-03 10:24:16 +01:00

Owner

We need to investigate issue we have with our pleroma. Seems like db issue but we need to investigate

muppeth added this to the 11.22 - November milestone 2022-11-03 10:24:16 +01:00

muppeth added the

Urgent!

🤔️ Investigate

Akkoma

labels 2022-11-03 10:24:16 +01:00

muppeth commented

2022-11-09 21:15:51 +01:00

Author

Owner

The most important thing is that most likely the migration caused to reload db schema probably adding missing or broken indexes which kicked the instance back to life.
I have done some minor db config tweaks which should help in the future. I also run some db maintenance tasks which we should run on regular basis eg. removing old remote content (older then 90days by default) etc. This helps keep the db in sane space.

With deployment of Roos we should upgrade current postgres from 12 to at least 13

We have managed to get pleroma or shall I say now Akkoma back online and kicking. Mainly due to migration to pleroma's fork called Akkoma which seems like a good choice at this point. It is more maintained with additional features and on good terms with upstream pleroma. Also many instances have moved there so its a good sign. The most important thing is that most likely the migration caused to reload db schema probably adding missing or broken indexes which kicked the instance back to life. I have done some minor db config tweaks which should help in the future. I also run some db maintenance tasks which we should run on regular basis eg. removing old remote content (older then 90days by default) etc. This helps keep the db in sane space. With deployment of Roos we should upgrade current postgres from 12 to at least 13

muppeth closed this issue

2022-11-09 21:15:51 +01:00

muppeth referenced this issue

2022-11-11 22:27:21 +01:00

[Pleroma] - Performance issues #361