this post was submitted on 15 Apr 2024
761 points (96.7% liked)
Memes
45665 readers
974 users here now
Rules:
- Be civil and nice.
- Try not to excessively repost, as a rule of thumb, wait at least 2 months to do it if you have to.
founded 5 years ago
MODERATORS
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
I really don't think we want to get into an uptime competition with Reddit lol
Every time one instance is down, I switch to another. I've never seen it down completely.
...you were saying?
my experience is a bit different. when the instance with your account and subscribed channels is down, it's just not a good experience
!? Future feature request for Lemmies : Backup and restore subscriptions to hop fast to another instance in case of downtime.
Sync has this
coolio 👌😎
I was just thinking this should be a "First time?" meme instead. It feels like there's always one instance down.
But it's nice that Lemmy as a whole is never down, just individual pieces.
I think lemmy.world beats reddit. And it's one of the worst offenders.
Across all the federated nodes? I'd take that competition, that's the point of the fediverse after all
What’s really cool is that content isn’t lost. As soon as those nodes come back, they get to fast-forward through all those queued updates.
What happens if the instance which hosts the community is down but other instances are online? I just made a post to a community on an instance that doesn’t exist anymore, but will other instances get that post, or is it reliant on the instance which hosts the community coming back online?
You may think of the community as down in that case, but overall the service is more resilient than a single community.
That post will only exist on your instance. Federation out to other instances would have to happen via the instance that community is on.
To be fair though, the chance that every Lemmy instance goes down at the same time is so much lower than Reddit going down. Sure, my instance might be unavailable, but I'd be able to hop onto the next one and continue.
Considering the distributed nature of the fediverse, the only way I could imagine the entire thing going down is either a botched update that everyone somehow manages to install at once, or a very sophisticated distributed attack.
The latter would be the most realistic as someone could find a way to poison the ActivityPub protocol or there’s a bug in it and then a message would go to all servers simultaneously.
Unless every instance goes down we will just call it a partial outage.