It’s a challenge, for sure. It is known that there are some inefficiencies in the codebase, which are actively being worked on. But besides that, it’s tricky to know where bottlenecks are until the user influx happens, particularly with the novel federation architecture. Maybe it’s impossible to scale, maybe not, but we only now are seeing a testable use case. I would expect optimization work to start bearing fruit, but these thing take time.
I totally respect this being potentially a big ask, but does anyone have a TL;DR of what caused or was the fix for the federation issue(s)? I don’t have capacity at this moment to look through Github Issues and PRs, but I’m curious