this post was submitted on 11 Jun 2023
16 points (94.4% liked)
Asklemmy
43952 readers
676 users here now
A loosely moderated place to ask open-ended questions
Search asklemmy ๐
If your post meets the following criteria, it's welcome here!
- Open-ended question
- Not offensive: at this point, we do not have the bandwidth to moderate overtly political discussions. Assume best intent and be excellent to each other.
- Not regarding using or support for Lemmy: context, see the list of support communities and tools for finding communities below
- Not ad nauseam inducing: please make sure it is a question that would be new to most members
- An actual topic of discussion
Looking for support?
Looking for a community?
- Lemmyverse: community search
- sub.rehab: maps old subreddits to fediverse options, marks official as such
- !lemmy411@lemmy.ca: a community for finding communities
~Icon~ ~by~ ~@Double_A@discuss.tchncs.de~
founded 5 years ago
MODERATORS
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
How much headroom do you have left on that? I'm considering starting up a public instance and would love to get an estimate for per-user workload on a federated instance.
With just me on the system, CPU is barely ever over 2 -3%. Load average looks good. Memory usage looks fine. You know what? Let me post some graphs for the past 24 hours, which, I've pretty much been on here for 24 hours straight. Again, I'm the only user on my instance, and this is all running in docker containers.
I've mentioned this in a few other threads, but I'm tempted to fire up jmeter and push some load through my instance just to see how it behaves if I slam the system via the API. I just don't feel like learning the internal API endpoints and all that right at this time though.
Super cool, thanks
Awesome, this is super helpful! I'd be using a very similar setup. It might be best to start small, invite a couple people on, and see how that memory scales. I'll be avoiding any auto-scaling unless it becomes a much bigger project.
Well, ideally each service would have their own dedicated resources to begin with. But, given all of the lemmy services + Postgres are running on 2 cores with 2GB of RAM, that's pretty impressive.
Anyway, autoscaling doesn't necessarily solve scaling issues without a lot of thought and planning. It's not always as simple as throwing more hardware at the problem, as I'm sure you already know.