High memory consumption on front-end server over time

[Bentroen]

The Next.js frontend deployment consumes about ~600MB RAM when starting, but rises to almost 900MB over a few days.

This smells like a memory leak somewhere - the backend memory consumption stays relatively the same in comparison.

This is particularly concerning because memory consumption accounts for >90% of our Railway bill :) so being smart about it could mean a better use of our contributors' generous donations.

Of course, the best approach would be to investigate why this happens and mitigate it. But a possible workaround would be to restart the service periodically. From this thread in Railway support, this is not a feature currently offered by the service, but allegedly this Railway template allows this possibility by offering scheduled restarts via cron.

It may also be worth exploring using more efficient, alternative JavaScript runtimes, such as Bun -- some people have reported shorter build times and overall better resource usage with it.

[OctoFlareDev]

🚰

[tomast1337]

I've created the branch https://github.com/OpenNBS/NoteBlockWorld/tree/feature/bun for testing Bun. While the project runs normally locally, I'm unsure if this is the right solution

[tomast1337]

With a questionable methodology, I spun up the project, entered every page in the Next.js application, and recorded the memory usage. This was done in development mode, so there may be some overhead.