Day 2026-10-18: report, CONTEXT; old-restic check not run yet (demo-hp's clean-up window opens after the 2026-10-19 04:15 CEST run), STATUS time corrected

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_0159rPz1ZhFKsS53msqPYxtS

documentation/audits/day-2026-10-18/P8/cleanup-window-not-yet.txt

@@ -0,0 +1,20 @@
+# P8 — old-restic check: NOT RUN (the clean-up window has not opened). Read 2026-10-18T14:33:43Z.
+
+## demo-hp's clean-up windows (hub DB with its WAL, one tar)
+id|customer|opened_at|closes_by|closed_at|box_result
+9|demo-hp|2026-10-12 02:16:53|2026-10-12 02:36:53|2026-10-12 02:17:06|pruned
+6|demo-hp|2026-10-05 05:25:53|2026-10-05 05:45:53|2026-10-05 05:26:04|pruned
+2|demo-hp|2026-10-04 05:36:14|2026-10-04 05:56:14|2026-10-04 05:36:17|nothing
+
+## The rule (hub/internal/offsitekeys/service.go:255)
+	windowCadence     = 6*24*time.Hour + 12*time.Hour // "weekly", with slack for the night chain's drift
+Last window 2026-10-12 02:16:53 UTC + 6 d 12 h = 2026-10-18 14:16:53 UTC (the STATUS time).
+The box asks for a window only AFTER its nightly off-site run (controller backup/offbox_window.go header).
+
+## demo-hp guest 9201 controller (docker logs since 06:00 UTC, [offbox] lines + scheduler)
+felhom-controller gitea.dooplex.hu/admin/felhom-controller:0.330.0 Up 3 hours (healthy)
+2026/10/18 11:09:36 scheduler.go:150: [INFO] [scheduler] Daily job offbox-backup scheduled for 2026-10-19 04:15 CEST
+2026/10/18 11:09:36 scheduler.go:85: [DEBUG] [scheduler] daily job registered: name="offbox-backup" schedule="04:15" nextRun=2026-10-19T04:15:00+02:00 totalJobs=15
+2026/10/18 11:09:36 scheduler.go:85: [DEBUG] [scheduler] daily job offbox-backup: next run at 2026-10-19 04:15:00 CEST (waiting 15h5m23s)
+
+Next run: 2026-10-19 04:15 CEST = 02:15 UTC; the window opens right after it. No window was asked for or granted by CC.