For a big instance getting an over-sized user base … roughly how many would that be and what could the instance do about it? I’d imagine a number of infrastructural things could be done before the core lemmy code base and design needs to be substantially changed or redesigned. Big separate database service, big beefy primary server/instance or even a cluster like kubernetes (which is what mastodon.social use AFAIK).
As for the alternative where many users are distributed more even across many instances, how well would that or can that scale with all of the community data that would need to be synced up between all the instances? From what I’ve gathered, it’s precise this kind of work that’s plagued lemmy.world somewhat and caused some of the issues that users have been having, largely, it seems, from the server being overloaded with “federation workers” timing out.
I have the technical knowledge to back it up, and I confirm your understanding is spot on.
Thanks!
Can we pick your brain on this?
For a big instance getting an over-sized user base … roughly how many would that be and what could the instance do about it? I’d imagine a number of infrastructural things could be done before the core lemmy code base and design needs to be substantially changed or redesigned. Big separate database service, big beefy primary server/instance or even a cluster like kubernetes (which is what mastodon.social use AFAIK).
As for the alternative where many users are distributed more even across many instances, how well would that or can that scale with all of the community data that would need to be synced up between all the instances? From what I’ve gathered, it’s precise this kind of work that’s plagued lemmy.world somewhat and caused some of the issues that users have been having, largely, it seems, from the server being overloaded with “federation workers” timing out.