How Large Should Servers Be and How Can That Influence the Federation Health?
/2024-09/session/6-a/
Convener: nigini (@nigini@social.coop)
Participants who chose to record their names here:
- Ruud (@ruud@mastodon.world)
- Tommi (@tommi@pan.rent)
Notes
Rudd: running two large servers… should servers be “that big”?
- One instance of Mastodon and one Lemmy.
- The size of the server doesn’t matter, as long as the load is distributed!!!
- It probably is: 90% of the users don’t care about the server they join!
- It would be great if we had a tool to move a whole community.
nigini: is it a problem to even use the “pick your server” as a UX.
tommi: the opposite position… I have a small community and that needs more users…
- We could create systems to decentralize.
- Maybe the server’s admin could send automatic messages… consider migrating.
- Also, how to deal with the “Threadification” of the fediverse!
- How to deal with big actors coming up?
Ruud: Threads is also an opportunity, because people are exposed to other “name@server”
- Mastodon.social agrees that their size is not ideal.
- The problem is: can you trust on the reliability of the the other servers?
nigini: reminder about LOLA proposal (better migration capabilities).
Very few people have activated federation in Threads.
- But: Threads are still not fully federate.
nigini: what are the signals that would give a good sense about the health of a server?
Rudd: size doesn’t give us that… I saw very large servers go down!
- Who runs a server? how they run a server! how to track that?
- Should we have vaults for admins (envelop procedures)
- Users side?
- More tooling: the server has data to know what would be aligned servers to them!
- Also, aligning on the covenants level.