Most of the communities I’m interested in are on lemmy.world, programming.dev, etc where they have a large activity there, but the UI/UX of kbin is what’s holding me here.
However, I feel like not all upvotes downvotes and even comments are federated to kbin, so I never get to experience the real hype of certain posts.
I’m wondering where the bottleneck is? Is it a certain i stance not pushibg to kbin.social, or is it kbin.social not being able to handle this huge amount of streams?
Also, if it’s the latter, would I have a better experience if I were to host kbin myself?
Currently, kbin has no delays in the queues. This may be due to not all streams being sent to us or an error in the federation. Over the next few weeks, this process will be improved.
The federation between the Mastodon instances I am using for testing is either immediate or takes a maximum of a few seconds. However, other instances may experience their own delays.
Wow, thx. It’s nice to hear an official statement.
Are you guys in contact with each other to investigate the issue?
I’ve noticed that some communities will be in sync and up to date with the /kbin magazine, but some are hours, even weeks, behind and I don’t know what the difference is. This discrepancy is present even on the same instance like lemmynsfw, where one community is up to date, but another community hasn’t synced in weeks.
If federation updates were missed e.g. during the DDOS attacks, on Lemmy or just when instances were creaking from reditMigration demand, these updates won’t be resent. Instances never activity seek out missing information to resynchronize.
…basically if it has been weeks, it’s just not going to happen.
Also worth noting downvotes don’t federate they are only know to each local instance.
My perception is that it is all working pretty smoothly now. I quite often switch between my Mastodon account and kbin and they always seem to be aligned.
Its not that one post is weeks behind and just got missed in a federation update. Its that zero posts have been federated in weeks. Its not a skipped record, its entirely halted.
That’s very much not what the post I replied to describes. I quote:
but some are hours, even weeks, behind
What you are describing sounds more like you are looking at a community/magazine which hasn’t been subscribed to from the other instance so, by design, posts don’t clone. This is expected and intended behaviour.
I think you misread communities as posts. The entire communities are up to weeks behind, not just certain posts from them
and no we’re not talking about unsubscribed communities.
There were talks of lemmy.ml limiting kbin but I don’t know if that was ever verified. There are other kbin instances you could try and see.
I use apps so lemmy experience is ahaping up to be pretty good for me.
It was verified and fixed, so users on kbin. social should at least be able to see their posts now https://kbin.social/m/kbinMeta/t/173366/lemmy-ml-is-no-longer-shadowbanning-kbin
We on kbin could always see lemmy.ml, it was their server which was accidentally misconfigured to reject information originating from kbin. As you say it has now been resolved.
.ml unblocked the kbinBot UA a couple days ago so as new posts and comments are made it should be syncing up. there are still a few instances out there that seem to not federate as well (in particular beehaw doesn’t seem to be federating beyond community discovery) but i believe that’s just some reverse proxy misconfiguration (the lemmy-ansible nginx config had some federation related fixes with the release of 0.18.1 they may not have applied) rather than anything intentional