feat(deploy): continuous WAL archiving to S3 for point-in-time recovery

Add pgBackRest-based PITR for the production database, shipped disarmed
(archive_mode and the base-backup timer gated off) so it stays inert until the
operator arms it before real payments — an un-armed deploy cannot pile WAL onto
the disk.

- Postgres now builds from a thin image carrying pgBackRest (archive_command runs
  in-process); the prod overlay wires an encrypted (AES-256-CBC), path-style S3
  repository with 30-day retention and a daily full base-backup systemd timer.
- Repository config/secrets flow through the PROD_PGBACKREST_* Gitea set and
  write-prod-env.sh; prod-rollback re-renders the same env so a rollback cannot
  disarm archiving.
- Grafana alerts watch pg_stat_archiver (failing / stalled), absent-safe on the
  test contour, which never archives.
- Fix the pre-migration pg_dump to snapshot the whole database (was backend-only,
  silently excluding payments).
- Document the PITR runbook, the arming sequence and the restore drill in
  deploy/README.md; record the measured cost/perf assessment.

deploy/ansible/group_vars/all.yml

@@ -27,3 +27,12 @@ docker_log_max_file: "3"
 # path (a slow service beats a killed one). Set swap_size to "0" to skip provisioning.
 swap_size: "1G"
 swap_swappiness: 10
+
+# Point-in-time recovery (pgBackRest). The base-backup systemd timer on the main host is
+# gated by pitr_enabled so provisioning stays inert until archiving is armed — the timer
+# would fail nightly if it ran before the S3 repository + stanza exist. Arm it as part of
+# the PITR rollout (deploy/README.md): flip it on with `-e pitr_enabled=true` (or a main
+# host_var) once the stanza is created. Only the `main` role reads these.
+pitr_enabled: false
+# systemd OnCalendar for the daily full base backup — a low-traffic hour on the main host.
+pitr_backup_oncalendar: "*-*-* 04:00:00"