The long awaited update of 0.19 is here so I upgraded. This gave me huge problems because the CPU load is constantly going up to 100 which brings my whole server down, not just Lemmy. There is a bug for it https://github.com/LemmyNet/lemmy/issues/4334 but it’s not getting the traction I thought it would get since very many instances have this problem.

I tried to contain the problem in the docker instance, but somehow limiting load to 0.5 in docker-compose didn’t do anything. Then further down in the bug in a comment someone suggested to do database.pool_size: 30 in lemmy.hjson because the default of 95 was to high. With then I got down the load to around 2 which is ok.

But with this higher load I had to shut down Nextcloud because that would then bring everything down together. My guess is that I didn’t really optimize Nextcloud and I have so many services running that it’s on the brink of not being able to run all of them together.

It seems Nextcloud is a heavy weight, and Lemmy wasn’t, but now something is wrong and it hogs resources. With the change at least I still can run Lemmy, need to find a similar solution for Nextcloud too.