Lemmy.world: why not focus on stability instead of growth?
This instance happens ti be down quite often.
Why do the admin doesn't close the subscription and focus to keep the instance stable? It's not like there will be blocked outside Lemmy, there ar plenty of instances out there.
From outside (and IMHO) it seems a power hungry needs (sorry,bI don't know if this does make sense in English).
This is a question (that I hope will flow to a good conversation), NOT an attack.
From my understanding, the stability issues drive from repeated DDoS attacks. I’m not sure locking new membership solves that problem but does run the risk of starving new members from joining and risking a shrinkage of content.
I’m sure the admins are focused on more than one thing at a time and closing the site doesn’t seem like it makes a whole lot of sense.
Since other instances seem to have taken action that suits you better, you can always move your account to one of them. You could still access communities hosted on .world, but you wouldn’t be inconvenienced by DDoS attacks taking down the server.
The operators of this instance are legit professionals and work diligently at resolving stability issues. The occasional problems with this instance aren't caused by resource or capacity problems, but by external forces (coordinated DDOS attacks) and inherent Lemmy issues. The reason smaller instances aren't targeted more is because terrorism tends to focus on affecting the most number of people possible (big instances are big targets).
Stopping growth won’t make anything more stable. The problems Lemmy.world seem to be experiencing are less related to load on the servers, and is more related to bad actors attempting to harm the instance. Closing signups won’t magically make the bad actors go away
The downtime is not related to user growth.
We do however plan to change our registration page, so it will refer to other instances and explain Federation in short. New users will be advised to choose other instances but will still have the option to sign-up here if they would like.
But all the troubleshooting these downtimes hasn't left a lot of time to work on this...
Hi Ruud,
I know my original post wasn't so nice, wasn't ment to be, I've written in different moment and that's the result. Sorry.
Technically speaking, isn't the Cloudflare protection efficient against the DDoS of the last days? What is overloading the server?
No problem, it's good to ask those questions. Sometimes we also learn from them.
The outages of last days were because of database queries that took very long, which were fired many times.
This is being looked into.