A few weeks ago Lemmy was buggy on computers and there were no good mobile clients out there, now on PC the site is pretty stable and fast, and there are now some pretty good iOS/Android clients too. Thanks to all the people who made this possible!
A few weeks ago Lemmy was buggy on computers and there were no good mobile clients out there, now on PC the site is pretty stable and fast, and there are now some pretty good iOS/Android clients too. Thanks to all the people who made this possible!
All of your points only considers the community itself, which is not my argument. And that can already achieved by just doing Vertical scaling like most instance currently do, I think even lemmy.world just using 32 core now (google cloud can possibly run with more than 200 vCPU).
I’m mostly approach this from technical standpoint. It’s impossible to have 100% uptime if there’s no horizontal scaling capability. For example on updating version, currently the instance will need to shut down for maintenance until it’s finished and usually there’s still some issue to fix. If horizontal scaling exist, the instance can update server (or add additional one), move the traffic a bit to test it, and then fully rollout if everything going well.
Not to mention a hardware failure, which could take a couple of hours to fix at least, some mental health communities should stay online at all time, someone mentioned there is research showing when a person is suicidal there are a couple of hours he is vulnerable, and there is some research showing online support can improve mental health.