the-custodian/docs/state-hub-railiance01-cutover-phase5-2026-07-06.md

T05 Phase 5 Evidence — Stabilization Window (2026-07-06)

Workplan: CUST-WP-0054-T05
Parent runbook: docs/state-hub-railiance01-cutover-plan.md
no_secret_material_recorded: true

Window

Field	Value
Started	2026-07-06 ~17:35 UTC
Target close	2026-07-09 17:35 UTC (72h)
Owner	custodian
Rollback	coulombcore state-hub Deployment + dumps retained ≥7 days

Phase 5 status: in progress.

Kickoff baseline (2026-07-06T17:41Z)

Check	Result	Evidence
Hub /state/health	pass	{"status":"ok","db":"connected"} via state-hub-primary
Row-count gate	pass	workstreams 640, tasks 4002, topics 14 (matches Phase 2 gate)
railiance01 pod	pass	state-hub-76fdffdc9d-dtvpb Running, 0 restarts (rev 7 @ main-375961c)
CNPG state-hub-db	pass	Cluster in healthy state, 1/1
coulombcore rollback	pass	state-hub replicas=0 (not deleted)
activity-core hub path	pass	in-cluster http://state-hub.state-hub.svc.cluster.local:8000; bridge 0
Consistency sweeps	pass	6/6 recent events exit_code=0, skipped_missing=0 (post-17:15 UTC)
Sweep writeback host	pass	repos resolve /home/tegwick/* (Phase 4)
Workstation tunnel	pass	state-hub-primary connected + LIVE → railiance01
Daily triage (post-cutover)	pass	manual trigger 2026-07-06T17:50Z; event a99e644e, report daily-triage-2026-07-06-67685f1c.md
issue-core emission	unchanged	issue-core-coulombcore tunnel connected; null-sink audit mode (Wave 4)

Kickoff sweep (source=phase5-kickoff): exit_code=0, 10 repos processed, 0 missing.

Day-1 check (2026-07-06T22:35Z)

tools/phase5-stabilization-check.sh — all green:

Check	Result
/state/health + totals	pass — 640 workstreams / 4002 tasks / 14 topics
railiance01 pod	pass — state-hub-76fdffdc9d-dtvpb, 0 restarts, age ~4h
CNPG	pass — healthy 1/1
coulombcore rollback	pass — replicas=0
activity-core	pass — in-cluster URL; bridge replicas=0
Sweeps (last 6)	pass — exit_code=0, proc=11, missing=0
Daily triage	pass — last run 2026-07-06T17:50Z
Tunnels	pass — state-hub-primary connected + LIVE

Window closes 2026-07-09 ~17:35 UTC (~65h remaining at check time).

Scheduled checks (activity-core)

Definition	Schedule	Event type
phase5-stabilization-daily (f3a8c2e1…)	0 9 * * * Europe/Berlin	phase5_stabilization_check
phase5-stabilization-closeout (e7d2b5a8…)	once 2026-07-09T17:35Z	phase5_stabilization_closeout

Hub-visible gates: health, totals (640/4002/14), last 6 sweeps, daily triage freshness ≤36h. Closeout fails visibly when any gate is red; passes record operator_signoff_needed for coulombcore teardown approval.

Monitoring

Run daily (or after incidents):

~/the-custodian/tools/phase5-stabilization-check.sh

Sweep history (24h):

cd ~/state-hub
uv run python scripts/compare_consistency_sweep_parallel.py --since-hours 24

Exit gates (T05 done)

From docs/state-hub-railiance01-cutover-plan.md verification checklist:

• /state/summary totals match pre-cutover (±0)
• activity-core daily triage succeeds once post-cutover (manual 2026-07-06T17:50Z)
• activity-core consistency sweeps succeed without fleet-state-hub-coulombcore
• Workstation :8000 serves railiance01 hub
• Writeback lands on railiance01 clone path
• coulombcore hub scaled to 0
• Cutover progress notes recorded (Phases 0–5)

Close Phase 5 / mark T05 done when: 72h window passes with no data-loss signals, sweeps stay exit_code=0 (lock_skipped acceptable), one post-cutover daily triage succeeds, and operator approves coulombcore hub teardown window.

Known watch items

1. Sweep budget — 300s processes ~8–10 dirty repos per tick; expect gradual reconciliation across 15m cron cycles (not a blocker).
2. Image tag — resolved 2026-07-06T18:15Z: Forgejo CI image.yaml #26 published main-375961c; Helm rev 7 deployed; post-upgrade sweep exit_code=0.
3. Stale tunnel — state-hub-coulombcore shows [STALE]; safe to retire after stabilization if unused.
4. issue-core — still coulombcore-chained; tracked under Wave 4 / T06, not T05.

References

• Phase 4: docs/state-hub-railiance01-cutover-phase4-2026-07-06.md
• Sweep runbook: state-hub/docs/consistency-sweep-runbook.md