maltfieldJun 12, 2023

How is lemmyworld so stable?

https://monero.house/post/5176

How is lemmyworld so stable? - monero.house

At the time of writing, Lemmyworld has the second highest number of active users (compared to all lemmy instances) * https://lemmy.fediverse.observer/list [https://lemmy.fediverse.observer/list] Also at the time of writing, Lemmyworld has >99% uptime. By comparison, other lemmy instances with as many users as Lemmyworld keep going down. What optimizations has Lemmyworld made to their hosting configuration that has made it more resilient than other instances’ hosting configurations?

Show threadpsychothumbsJun 12, 2023
Looks like the guy who runs it runs a lot of fediverse servers, I guess he knows what he's doing: https://lemmy.world/u/ruud
Show threadmaltfield
Yes. And I'm asking him to share his tweaks here with the community so that others instance admins can shore-up their servers :)
Jun 12, 2023 at 6:08pmWeb
Show threadPriorProjectJun 12, 2023

Fwiw, he has been providing quite a lot of transparency in his posts to this community. He's shared his hardware config in detail, posted maintenance posts with brief descriptions of what he's doing, and replied to comments around specific config tweaks. I haven't catalogued a list of links, but I've seen him do all of these things in the last 48h. It's easy to imagine that all these things could be compiled in real time into a how-to, but it's a pretty big deal just to keep the lights on right now, and pretty difficult to understand whether tweaks that helped your setup are generally applicable or only situationally useful and happen to perform well for your specific setup.

I'm sure we will see more high-performance Lemmy guides in the future, but at this point no one has more than 36h of experience with high-performance Lemmy. Give them a minute to catch up.

Show threadKrisJun 12, 2023
More like throw more $$$ at the problem has been the solution.
Show threadMetaCubedJun 13, 2023
Saying this without any knowledge of lemmy's backend... Large, high user count databases with (very quickly) growing demand take more power than generally expected. At a certain point, throwing money at the problem is the solution.