Instance admins can “purge” it, but it is only local to that specific instance.

https://en.wikipedia.org/wiki/Caucasian_race

yes they will be redownloaded on request.

You still don’t have a clue, but I am tired of trying to argue with obvious bad-faith Russian trolls like you.

Yes, restarting Lemmy somehow resets the memory use of the database as well.

Nothing particular, but there was a strange bug in previous versions that in combination with Lemmy caused a small memory leak.

Try switching to Postresql 16.2 or later.

Hmm, there is a memory leak issue with PostgreSQL versions lower than 16.2, so instances would run out of RAM and start swapping after a while. But it would surprise me if lemmy.ml specifically was suffering from that as the devs surely know about this issue.

You get focus time?

They need to give you some time to answer emails from management 🙄

Yes, I also noticed that. It used to work a while ago.

If you like our communities so much, you could also create an account on slrpnk.net 😊

But yeah, two smaller instances with not much overlap in interests by their users tend to not federate well and you need to go the extra mile to search out the communities and subscribe to them.

First of all this is mainly a feature to protect existing instances against spam from other badly managed instances.

But yes, admin approval scales surprisingly well if you have several admins per instance and many smallish instances, and it could be further improved by giving existing users secret invite links than don’t require admin approval (AFAIK also already supported by Pixelfed).

Those are largely useless against spammers, but indeed there could be other options like giving already registered users usage limited invite links etc.

Upgraded from 0.19.2 without problems. Thanks!

Cold outside and warm water makes it easy to convince them to change their minds.

Are there some details about using the new persistent queue in more complex setups? I remember initially it was supposed to run in a separate worker container or so?

I don’t agree that this is something needed, but if you decide to work on it anyway, maybe try to make it compatible with Pixelfed’s Sup. https://github.com/theSupApp

Webpush or UnifiedPush would be the answers, but neither is currently supported by Lemmy.

Edit: but please don’t abuse Lemmy’s DMs as a sort of instant messenger. It’s not meant for that and also insecure. It’s meant to be asyncroneous like email at most and you don’t need push notifications for that.

By default, admins should focus mostly on the activity inside their own instances. Let the behaviour of their users in other instances up to those admins;

This can’t possibly work due to two related reasons:

1. Small instances don’t have enough moderators to deal with a large number of remote users alone.
2. Only the threat of losing their main account on the home instance is a deterrent for some users to not troll on remote instances.

I think it should be rather that admins should leave moderation on their own instance mostly to the local community mods and specifically focus on what the users that they are responsible for are doing on remote instances. If admins are unwilling to do that, then this would be grounds for a defederation as this is basically the most important job to build a good reputation of an instance.

Looks like they are playing Frets On Fire 🤣