I started a private instance today, and maybe I’m being impatient but I’d like to know if there’s something I can do on my end to keep my instance more closely in sync with the remote communities I’ve subscribed to. I don’t seem to be missing any new posts, but there are so many missing comments which makes the sort options (Active, Hot) bring up posts that sometimes are weeks old. Are there any configuration options or tuning options that could help my instance “keep up”?
Active and Hot are kinda weird right now, they’re fixed in the next version, but your best way of seeing trending stuff currently is by doing Top Day/Top Week.
Some of the bigger communicaties like Lemmy.world and lemmy.ml are experiencing terrible growing pains from the massive expansion of Lemmy. You can help your own instance by adding more federation workers in the admin panel, assuming your hardware is up to the task, but I think a lot of the missing comments are a factor of the big communities being overloaded.
Increasing my federation workers to 1500 resolved a lot of my warnings in my logs, but I’m still missing tons of comments from Lemmy.world.
I’ll try that, thank you for the suggestion. You’ve also helped me confirm that my instance is federating correctly with lemmy.ml, kbin.social and your own instance as well, so I think we’re off to a good start.
The first posts you get from before when you started your instance are going to be empty.
The next ones you get from the communities you subscribe to are going to be normal.But I’ve found lemmy.ml also has issues keeping me up to date with new stuff, other instances work fine for the most part, but this one I’ve seen a post with 0 comments in my instance and the original already had had a whole conversation.
I see. I’ve a had a similar experience with some posts from beehaw.org and lemmy.world, but of course those are huge instances already and growing quickly.
Nothing you can do about it on your end. ActivityPub seems to require the larger instances to send outbound notifications, messages received more than 10 seconds old are tossed out, and it appears right now the larger instances can’t keep up with delivery in that time frame. Makes sense… lemmy.world instance has 3.6K users and 2200+ federated servers. If only 10% of the users upvote one post, that’s over 700K out going notifications already. I’d imagine a larger portion of the current users being more active considering large chunk of us are the more active Redditors seeking refuge…
That makes sense, thank you for the insight. Also it’s a good sign that this post and the comments are federated correctly from my instance to lemmy.ml, kbin.social and your own instance as well. Very cool to see this in action.
There is an Open Issue on Github
