Saving of current progress of generating chunks appears to be a pretty wasteful process.
This causes a lot more activity to occur on the main thread, for data that becomes irrelevant very shortly after.

All saving of proto chunks even helps with is resuming a chunk that was cancelled mid generation at server shutdown.

That's a pretty extreme use case, when you could simply not save them, and allow the generation to start over the next request.

I really recommend removing this, so only Level staged proto chunks are saved.

This gets rid of the entire method at the chunk scheduler that processes these, saving a ton of work.