Yeah, this latest batch of federation delays seems to be a series of bugs interacting with each other. However, the primary cause seems to be a bug in kbin in which it is flooding out tons and tons of activities (votes, downvotes, follows, etc.) repeatedly, seemingly stuck in a loop (relevant post from an .nz instance admin). The world admins put a filter in place on their firewall to block it a couple hours ago and the federation seems to be catching back up. I just checked and right now ani.social is 4.5 hours behind world, down from a peak of 11.5. Due to a different, unrelated lemmy issue, small increases in latency to the world instance server (located in Finland) can cause much slower federation (relevant github issue). So, instances located close by in the EU seem to have been able to handle this better than other regions.
Yeah, this latest batch of federation delays seems to be a series of bugs interacting with each other. However, the primary cause seems to be a bug in kbin in which it is flooding out tons and tons of activities (votes, downvotes, follows, etc.) repeatedly, seemingly stuck in a loop (relevant post from an .nz instance admin). The world admins put a filter in place on their firewall to block it a couple hours ago and the federation seems to be catching back up. I just checked and right now ani.social is 4.5 hours behind world, down from a peak of 11.5. Due to a different, unrelated lemmy issue, small increases in latency to the world instance server (located in Finland) can cause much slower federation (relevant github issue). So, instances located close by in the EU seem to have been able to handle this better than other regions.