@argv_minus_one Is that scaling problem a software issue, or a hosting issue? There are other Fediverse platforms like Akkoma that use Elixir, so maybe they’d fair better? Could also pick several federated instances to distribute users to.
Lemmy is written in async Rust. The language isn’t going to create a scaling problem. Well-written async Rust applications have handled vastly heavier workloads than Lemmy without a hitch.
There are, however, some serious performance bottlenecks that need to be dealt with, and it remains to be seen whether any more bottlenecks remain undiscovered in either the protocol or the implementation. To be honest, as someone working on a Rust+Postgres application myself, this is the sort of thing that keeps me up at night.
Hosting can of course be an issue as well. I’m under the impression that Beehaw had to go up several tiers in its hosting plan in the last few days in response to the surge in demand. I assume this was done to work around the aforementioned bottlenecks by simply throwing more hardware at the problem, but I don’t know.
@argv_minus_one I see. The more I look into it, the more I think Lemmy should still be considered beta software like kbin, TBH. Some important features are still missing and the optimization is lacking.
@argv_minus_one Is that scaling problem a software issue, or a hosting issue? There are other Fediverse platforms like Akkoma that use Elixir, so maybe they’d fair better? Could also pick several federated instances to distribute users to.
Lemmy is written in async Rust. The language isn’t going to create a scaling problem. Well-written async Rust applications have handled vastly heavier workloads than Lemmy without a hitch.
There are, however, some serious performance bottlenecks that need to be dealt with, and it remains to be seen whether any more bottlenecks remain undiscovered in either the protocol or the implementation. To be honest, as someone working on a Rust+Postgres application myself, this is the sort of thing that keeps me up at night.
Hosting can of course be an issue as well. I’m under the impression that Beehaw had to go up several tiers in its hosting plan in the last few days in response to the surge in demand. I assume this was done to work around the aforementioned bottlenecks by simply throwing more hardware at the problem, but I don’t know.
@argv_minus_one I see. The more I look into it, the more I think Lemmy should still be considered beta software like kbin, TBH. Some important features are still missing and the optimization is lacking.
I do consider Lemmy to be beta software.
But it’s currently the best option.