jo,

So maybe instead of more forks that focus on rolling out new features, could the devs of Foundkey, , , et al work on your architecture so that even a relatively small (in comparison to some of the Misskey and Mastodon servers) 1500 user server, let alone a 4000 user one can function without falling over, suffering extended timeouts, having server costs blow out thanks to database performance issues, and then decimate it's community when folks are forced to migrate, and then those folks struggle to get their data exported.

So much front end work and not enough backend is the present legacy of the pre-Misskey v 13 forks.

Marie, (edited )

@jo One thing I can't understand is Iceshrimp's recent approach at let's throw searching onto the same Postgres.

While for small instances this is fine for bigger ones this is just a bottleneck cause using Postgres is more intensive the bigger the DB in total for FTS(Full Text Search) than Meilisearch and co as it has to do more behind the scenes than the optimized search engine DBs.

Meilisearch and Supabase actually have great articles explaining where using Postgres is a major downside or not.

https://blog.meilisearch.com/postgres-full-text-search-limitations/
https://supabase.com/blog/postgres-full-text-search-vs-the-rest

jo,

@Marie I think a big part of the problem with Misskey forks is that most devs are all on smallish instances and are basically doing things with their own instances in mind - mostly feature adding - and not really thinking much how will this scale or impede performance at scale. And tbh, given some of the stuff that has been raised at times you'd have to think some of the IS team think small and niche anti-social-media is their goal.

None of the fediverse safety stuff that Kaity had raised and was supposed to be a key part of the fork focus has ever been picked up by the team. And our move to IS to find it performing just as terribly , has left no time or energy to be able to re-add our Hajkey features, let alone contribute to the fork. So you can see why we've thrown up our hands and are just going to work with something close to Misskey upstream and hope that we finally get some stability back and rebuild something of community we once had. The last five months have been an absolute exhausting & disillusioning clusterfuck, in stark contrast to the first seven.

testing,

@jo you prefer a keychain over a bunch of keys ​:blobcat_laughing:​

  • All
  • Subscribed
  • Moderated
  • Favorites
  • random
  • slotface
  • kavyap
  • thenastyranch
  • everett
  • tacticalgear
  • rosin
  • Durango
  • DreamBathrooms
  • mdbf
  • magazineikmin
  • InstantRegret
  • Youngstown
  • khanakhh
  • ethstaker
  • megavids
  • ngwrru68w68
  • cisconetworking
  • modclub
  • tester
  • osvaldo12
  • cubers
  • GTA5RPClips
  • normalnudes
  • Leos
  • provamag3
  • anitta
  • JUstTest
  • lostlight
  • All magazines