You better not write your API in Python, or any language/library that uses amortised algorithms in the standard (like Rust and C++ do). And let's not mention garbage collection.

Huh? Could you please clarify wrt to Rust and C++? Can't they use another allocator if needed? Or that's not the problem?

I’m asking because I genuinely don’t know - what are “pages” here?

That’s a fair question. A page is the smallest allocatable unit of RAM, from the OS/kernel perspective. The size is set by the CPU, traditionally 4kB, but these days 8kB-4MB are also common.

Or you can set the vm.swappiness sysctl to 0.

If you’re writing services in anything higher level than C you’re leaking something somewhere that you probably have no idea exists and the runtime won’t ever touch again.

> I DON’T WANT THE KERNEL PRIORITIZING CACHE OVER NRU PAGES.

Then tell the Kernel about it. Don't remove a feature which might benefit other things running on your system.

If you're getting this far into the details of your memory usage, shouldn't you use mlock to actually lock in the parts of memory you need to stay there? Then you get to have three tiers of priority: pages you never want swapped, cache, then pages that haven't been used recently.

Can mlock be instructed to f.ex. "never swap pages from this pid"?

The application requests this itself from the Kernel. See https://man7.org/linux/man-pages/man2/mlock.2.html

From the link, mlockall with MCL_CURRENT | MCL_FUTURE

> Lock all pages which are currently mapped into the address space of the process.

> Lock all pages which will become mapped into the address space of the process in the future.