Cas is a user on cybre.space. You can follow them or interact with them if you have an account anywhere in the fediverse.
Cas @cassolotl

Some places are struggling with the results of Mastodon being popular, and are blocking instances to reduce the strain. :/ cybre.space/media/ob0Hwe1rkAAd

· Web · 2 · 2

@cassolotl i think this is ok? not every node needs to be connected to every other node

@bea Yeah, I agree! I just feel sad that some mutual follow links will be lost, I guess. What if they were good friends!

@bea @cassolotl :musical_note: The mast node's connected to the… witch node. The witch node's connected to the… oct node. The oct node's connected to the… cat node. The cat node's connected to the… mast node. :musical_note:

(To the tune of "Dem Bones")

@cassolotl It seems like part of this is an issue with how Mastodon issues affected GNU Social ones? I think there's some kind of bug that causes non-Mastodon servers to get hammered with requests repeatedly?

@maiyannah would know better than me. If there's an easy fix to be had here, it definitely needs to happen.

@deadsuperhero @cassolotl In that case, it's a single user-instance (well, like 2 people and a few bots) so load is going to be more pronounced.  That said, when a new instance networks with us, it easily spikes the load average very noticably.  I regularly had 30+ load average spikes on HLA whenever a new masto instance - even a smallish one - networked.

@maiyannah @cassolotl Any theories on what might be causing such spikes?

I'm curious as to how this could be minimized on Mastodon's end.

@deadsuperhero @cassolotl Looks like it tries to grab whole feeds at once.  Might be the conversation grabber being overzealous.  But I'm from the outside looking in, so I'm not sure.
@deadsuperhero @cassolotl Getting every notice for my account there is over 41k notices.

@deadsuperhero Oh interesting! Okay, thanks. :)