I used Fastify for serving backend (not static) content to the frontend of this web property and it has good performance. My problem is if I am serving a string (which could change), it shouldn't take 8ms to shunt it to the frontend especially on dev mode.
It wasn't a problem that I cared about too much, but I went to experiment with uWebSockets on serving that same data. 😄, can you believe it, freaking served the same string in 1 ms, so basically 87.5% improvement in latency. That's all I needed to know and went about building the same abstractions I had on Fastify on uWebSockets, so switching to it was very very easy.
I'm happy to have seen that the abstractions I built are sufficient so that I don't have to do major changes if I desire to switch web servers in the future. It's just the way to go to prevent lock in. I am deadly serious about high performing applications as it keeps costs down.
On the game development front is more of me taking a loose online course on Godot which is a really powerful free open source game engine. Epic Game's Unreal Engine is also free and that one is far more powerful, but I don't need all of that power. I'd like to create a Tyrian-like space shooter. Tyrian is a very old Dos/Win9x game that I still return to in the browser because 1. its a silly game, 2. its a fun game. You die, you restart the level and play the level again. You can up and downgrade your ship's guns, pick a different ship type and these options change as you go through the levels. You can collect data cubes that reveal some funny messages (transmissions) that kind of motivate you to collect them. The controls I feel were spot on. I'd like to make at least one level and put it on my phone as a modern Tyrian. If I find high motivation in bringing it to the Android, iPhone, Mac and Windows stores then I'll hire people to get that done.
Anyway, it's just a hobby not a business idea. There's no business motivation around the 2d/3d game development with Godot front right now though it wouldn't be a bad idea.