ci: faulthandler_timeout permanently in the hub-postgresql job

tsk-c6ckn5jhwxgp open

Oliver adopted this at the 2026-11-03 jour fixe (req-tmx2jadxrqvt).

Goal: when hub-postgresql hangs, the log shows every thread's stack at once instead of a silent 90-minute timeout. #1255 (diagnosis only, closed) showed the mechanism; #1253 added the 90-minute job timeout.

Set pytest's faulthandler_timeout for the hub-postgresql job, below the job's 90-minute timeout-minutes and well above the slowest single test, so a hang dumps stacks before GitHub cancels the job.
It costs nothing while nothing hangs; check that a green run's duration doesn't change.
Add a short context/process.md entry next to the job-timeout entry: why it's there, and why it's not pytest-timeout (that entry already rejects it).
Optional: same for hub-sqlite and hub-mariadb if it is the same one-liner.