monthly_maintenance: reschedule to 0 0 15 * * (15th at midnight)

Previously ran daily at 3am with internal self-gating. Now scheduled directly
on the 15th — uptime and interval gates still apply; a skip means next month's 15th.

Orchestrators/monthly_maintenance.sh

@@ -7,15 +7,14 @@
 #   1. Server uptime >= MONTHLY_UPTIME_THRESHOLD_DAYS days
 #   2. Last run was >= MONTHLY_RUN_INTERVAL_DAYS days ago (or never run)
 #
-# ── WHY UPTIME-TRIGGERED, NOT CRON ───────────────────────────────────────────────────────────
+# ── WHY UPTIME-GATED ─────────────────────────────────────────────────────────────────────────
 #   A scheduled reboot resets uptime. Monthly tasks (ZFS scrub, SMART long test) need a
 #   stable, settled system — not one that just rebooted. Uptime-gating ensures maintenance
 #   only runs after the server has been healthy for a full month, never immediately post-boot.
-#   The daily cron is just the trigger mechanism. The uptime and interval checks inside
-#   the script are what enforce the monthly cadence.
+#   If uptime or interval gate is not met on the 15th, the run is skipped until next month.
 #
 # ── HOW TO CALL ──────────────────────────────────────────────────────────────────────────────
-#   Cron: 0 3 * * * — daily 3am check. Script self-gates — calling it daily is safe.
+# Schedule: 0 0 15 * *  (15th of each month at midnight)
 #   Silent exit 0 when either gate is not met. Only outputs when maintenance actually fires.
 #
 # ── STATE FILE ────────────────────────────────────────────────────────────────────────────────
@@ -153,7 +152,7 @@ if [[ "$SHOW_STATUS" == true ]]; then
     fi
 
     echo ""
-    echo "  Schedule:   0 3 * * * (daily check — script self-gates)"
+    echo "  Schedule:   0 0 15 * * (15th of each month at midnight)"
     echo "  Force flag: --force bypasses both gates"
     echo "━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━"
     exit 0