I'm running all those services on the same Pi, and honestly, I gotta say.. not at all! Even when everyone's loading in at the same time at session start, it's snappy and responsive. It struggled a bit with 5e (5e's base system has little to no automation which means all modules need to redefine the codebase themselves. The result is a very broken, very laggy mess) but PF2e is insanely fast.

Even while doing automated rolls with complicated calculations and automated resistance values on damage taken, Node.js peaks out at maybe 30% CPU on a very complex roll, and idles at such a negligible amount it may as well be 0.

As for the modules, while I do think they're awesome, my suggestion is to play with things around in vanilla PF2e system, and add them one by one as you feel "hmm, maybe quality of life could be better on this" or "this could be spicier in terms of effects". Mostly because, while it's not too bad to follow the errors by using the web console, if you need to debug, it's easier to go "oh this thing that broke is regulated by module x" and knowing what each module does helps a lot, but also because the base Pathfinder 2e system actually already has a lot of quality of life and automation baked in and you probably shouldn't install things that the system might already automate and might end up conflicting and breaking things.

Anything prefixed by "Pathfinder 2e" and "PF2e" should be safe, but only really install them if you want the features they offer. If you come across anything while playing around with it, feel free to ask, or check stuff out in the Foundry Discord if you haven't, they've got some great resources in there for newcomers.