Hi, we are running Docker based Discourse 2.7.7 with a very large user base. After upgrading from 2.4.2 to 2.7.7, we are experiencing that some Postgres process are stuck at 100% CPU. Further investigation results that such queries seem to cause the problem: discourse_prod=# select pid, datid, que…

We’ve found the issue was caused by a custom SQL query in a user badge. So this is not an upstream issue. Thanks.

This query seems to come from Jobs::BadgeGrant.

I suspect that you’ll also want to make sure that you vacuum and reindex as described in PostgreSQL 13 update .

Postgres가 대규모 데이터베이스에서 CPU 100% 사용, Discourse 2.7.7

pfaffman (Jay Pfaffman) 8월 30, 2021, 1:11오후 5

PostgreSQL 13 업데이트에 설명된 대로 VACUUM과 재인덱싱을 수행해야 할 것으로 보입니다.

글		댓글	조회수
Postgres using 100% of CPU Self-hosting server-resources	16	20205	8월 25, 2018
Unusually high CPU usage Self-hosting	31	1414	2월 18, 2026
Some queries that don't finish Product Support	16	1099	4월 5, 2019
Long-Running Sidekiq Jobs Feature	21	2142	12월 24, 2020
Primary Postgres database process (postmaster) eating all CPU Self-hosting server-resources	13	4463	5월 28, 2020