Ideas for memory usage reduction

- [ ] Truly sleepy containers @phot0n
- [ ] Subinterpreters https://github.com/frappe/caffeine/issues/76 @aryanbaburajan
- [ ] Single workerpool process https://github.com/frappe/caffeine/issues/39
- [ ] HACK: Unload Python modules that aren't frequently used at runtime.
- [ ] Evaluate other memory allocators, python runtimes
- [ ] Profile and fix imports [just like before.](https://frappe.io/blog/engineering/reducing-memory-footprint-of-frappe-framework)

First add profiler integration: https://github.com/frappe/frappe/issues/35567

Start with the profiler integration referenced in frappe/frappe#35567, then read the linked caffeine issues for subinterpreters and a single workerpool process. The issue lists several possible memory-reduction directions but names no files, tests, or concrete completion criteria, so the scope needs to be narrowed before implementation.

Written by the indexing model from the issue text.