Francis 🚀 Gulotta@ceejbot @ummjackson I think speed vs perceived speed is the Crux. Super latent is fine if we don't know it's happening.
Jef Poskanzer@reconbot @ummjackson @ceejbot I think I ran into a case where speed vs. perceived speed becomes apparent. When I reply to a toot on another instance, an optimization makes my reply show up immediately; but if I reload, it's not there, and doesn't show up again until it comes back via federation. Which seems to take quite a few minutes.
@reconbot @ummjackson @ceejbot Actually this happens even to plain old local non-reply toots. They show up immediately due to the optimization, but if you reload they are not there until later. If someone favourites the toot, that can show up before the toot itself, which is interesting. Maybe notifications have a different queue?
@ceejbot @ummjackson @jef I think so. We have 3 queues at least.
https://github.com/tootsuite/mastodon/blob/master/Procfile
`-q default -q mailers -q push`
I can't find the job classes so I can only assume. ¯\_(ツ)_/¯
Ceejbot has a new jobYou're missing a queue! `-q pull` or federation will not work properly.