The last few days there appears to be quite a bit of delay between the synchronization of instances. Am I imagining that or is there something going on?
The last few days there appears to be quite a bit of delay between the synchronization of instances. Am I imagining that or is there something going on?
which instances?
some are just under-powered and cant handle the influx from big places like .world. then sometimes when theres downtime, it takes an extra amount of time to ‘catch up’
.world will also stop trying for 24 hours if it detects a sync failure (no connection, etc) for long enough. moist.catsweat.com is up to date.
.world is most noticeable but other instances also seem delayed. I made a post on progressivepolitics three hours ago and the only reaction I received was a .ml user.
Went over to check on .world and my post has not yet shown up there.
looks like something is up with lemmy.ml -> lemmy.world
https://grafana.lem.rocks/d/bdid38k9p0t1cf/federation-health-single-instance-overview?orgId=1&var-instance=lemmy.world&var-remote_instance=lemmy.ml&from=now-2d&to=now&timezone=browser
seems specific to your instance.
It looks like literally no one on lemmy.ml was subscribed to !progressivepolitics@lemmy.world, which is perhaps part or all of the problem? IDK. I just clicked “Subscribe”, and my subscription is “pending.”
I’ve always assumed that when my subscription says “pending,” that means no one else on my instance was previously subscribed, but maybe I’m wrong to think that 🤷
Edit to add: several minutes later and it’s still “pending,” which makes me think the problem goes deeper. I don’t sysadmin the instance so this is out of my hands.
It’s not normally about power level, but about how far apart the servers are. Because they’re sent serially and have to be acknowledged before they can be processed, it means that world can send out content faster than regionally distant instances can process them because of the geographical latency added to every event