Thoughts? Seems like we are one of the only bigger instances with open sign ups still as well.
Edit: didn’t mean to start any instance tribalism. Like others have said, its not a competition, and its better that users are spread across instances than piling into one. Just think it’s interesting to watch the different instances grow and change and see where people end up congregating.
lemmy.world has the IT Operations technical experience to keep their site up, the others need to catch up.
I’m developing some Kubernetes tools to make it easy for people to create their own scalable instances like lemmy.world has, so I’m hoping that that can make it a bit easier.
Sorry, let me clarify. My point was that I hope Ruud doesn’t have to start shelling out thousands of dollars because everyone decides to register there and he needs to keep increasing capacity.
Thing is, lemmy doesn’t support clustering/horizontal scaling. So there are limits to how much increasing you can do. You can beef up with a database cluster, add a separate reverse proxy, and increase the specs of the hardware lemmy is running on (but hardware can’t be expanded limitless), but that’s about it. Once you hit the limit of what a single instance of the lemmy software can handle, you cannot scale anymore. Pretty sure you will hit the limit long before you reach thousands of dollars.
Hopefully by then, Lemmy’s source is augmented to support HA/horizontal scaling
Yeah. But horizontal scaling (well horizontal scaling in a system like this where you need clustering so the instances talk to each other) is hard. And I think there are a lot of other things that need to be polished, added and worked on before that. It would probably also need somebody with knowledge of clustering to start contributing. I think step 1 needs to be that the dev team needs more help properly tuning the database use. The database is very inefficient, and they lack the skill to improve it:
So getting help improving the database is probably the #1 thing that can be done to deal with the scaling problem.
I fully agree - there’s no excuse why the DB should be falling over when I’ve seen a single postgres instance (with a read replica, granted) handle >1M users just fine.
Unfortunately my SQL skills haven’t improved since my DB class in university, so I won’t be much help. I’ll be keeping an eye on the repo of course, and I can give some consulting/guidance or even open some PRs myself when they decide to implement horizontal scaling.
Yeah. I’m in the same boat. My SQL skills aren’t impressive either since there are other people at work that handle optimization. Haven’t used rust either (yet) so cannot really contribute there either. Though I’m considering potentially starting work on a cross platform mobile app. I haven’t worked with mobile apps for a good six or seven years, so I feel like it’s high time I get back up to speed. (But knowing me, I’ll end up making something half finished and the start procrastinating)