Module resource 

Staying out of the way on a loaded production server.

Priority, not thread starvation, is the primary lever: a nice’d worker uses idle capacity and is preempted the instant Apache wants the core. Measured separately on 610 MB of real archives: FADV_DONTNEED takes page-cache residency from 610 MB to 0 MB at no cost in wall time or CPU (33.68 s vs 34.14 s, within noise), where zgrep -a, rg -za and rg -za -j1 all leave the full 610 MB resident and evict whatever Apache and MariaDB had warm.

The platform asymmetry is real and is not papered over. FADV_DONTNEED evicts pages after reading; macOS has no posix_fadvise at all and its nearest equivalent, fcntl(fd, F_NOCACHE, 1), prevents caching before reading. So there are two hooks — prepare at open and finish after the archive has been read — and each platform does its real work in a different one:

	Linux	macOS
prepare	FADV_SEQUENTIAL (readahead hint)	F_NOCACHE (the actual work)
finish	FADV_DONTNEED (the actual work)	no-op

Functions

finish

Called after an archive has been fully read.

load_average_1m

One-minute load average, from getloadavg(3): /proc/loadavg on Linux, the kernel’s own figure on macOS. None where there is none.

load_gate

Start adapting the concurrency to the load, or None when there is nothing to adapt: no --load-limit, a single job, or no load average to read.

next_gate

The gate after one more look at the load.

prepare

Called immediately after opening an archive, before it is read.

set_io_idle

Put our reads in the idle I/O class so they queue behind Apache’s.

set_priority

Lower our own scheduling priority. Unprivileged processes may always lower (never raise) their own, so this needs no capabilities and cannot fail in a way worth reporting.

starting_gate

The gate before any worker has started. All of the load is someone else’s at this point, so it is compared as it stands.