Mastodon is also an ultra-heavyweight in terms of compute resources it consumes per daily active user served. This is one reason (among multiple) that I would never run Mastodon as an ActivityPub microblogging instance.
Mastodon because it is the largest number of instances and users, but the principle is the same. It hasn’t happened with Lemmy yet because this subset of the Fediverse is ridiculously small compared to Mastodon/Pixelfed
ThorrJo@lemmy.sdf.org 1 week ago
rglullis@communick.news 1 week ago
Have you tried running any of the alternatives for more than a few hundred users?
Mastodon is bad for small and single-user instances because it wasn’t designed to serve so few people. Once it starts growing beyond a certain level of activity, most of your overhead is on sidekiq/redis side of things.
I’ve seen the Pleroma devs openly saying that their service can not handle more than 20k users, because they rely on the database to build the timeline and they don’t have a strong caching system.
So yeah, there are a bunch of tradeoffs and Mastodon is not trying to optimize for anything in particular. But once it reaches a certain size the operational cost (per user) does go down.
Blaze@lemmy.dbzer0.com 1 week ago
I’ve also heard that Lemmy is cheaper to run than Mastodon
jlai.lu/comment/13035055
rglullis@communick.news 1 week ago
You of all people here should know that the cost of running the service is not the real issue. Even if Mastodon takes 3x as much hardware to run as Lemmy, the cost of hardware is still pennies per user.
Blaze@lemmy.dbzer0.com 1 week ago
You are the one who brought the size of the platform as a reason for costs not being an issue here compared to Mastodon, so I assumed you meant that having more users would mean more costs for the admins?
Or do you mean that with more instances, there is a higher chance of some admin having issues covering the costs?
rglullis@communick.news 1 week ago
The latter.