Please implement vm.memory.size[dirty]

dirty = active write cache

On Linux it can be read like this: grep ^Dirty: /proc/meminfo

Result is in kilobytes.

Justification: I was frankly surprised agents don't read this. "Dirty" size, the area that is unwritten to disk yet from OS cache is a MAJOR performance metric. Especially that it is also used for CIFS and NFS caching. It's much more critical than many other vm.memory.sizes already implemented (e.g. wired, pinned, anon, file, exec, buffers have next to none practical utility to users).