state-hub/workplans/STATE-WP-0071-railiance01-workstream-caller-deploy.md

id	type	title	domain	repo	status	owner	topic_slug	planning_priority	planning_order	created	updated	state_hub_workstream_id
STATE-WP-0071	workplan	Railiance01 workstream caller migration deploy	infotech	state-hub	finished	codex	custodian	high	71	2026-07-09	2026-07-09	fb442549-8491-4a83-87d6-c2849dd85b30

STATE-WP-0071 — Railiance01 workstream caller migration deploy

Parent coordination: STATE-WP-0070 T01 (phase-2 legacy retirement — deploy and meter attribution).

Hub message: ops-warden inbox 467153f3-30ee-44c0-b89c-7f99fdf9f005 (2026-07-09).

Runbook: docs/evidence/railiance01-state-hub-deploy-20260709.md

Goal

Deploy workstream→workplan caller migrations to production paths so legacy-meter GET /workstreams/* traffic drops and usage is attributed to named components (state-hub.fix-consistency, activity-core resolvers).

Context

Two hosts participate; do not conflate them:

Host	Role
WSL workstation	State Hub API + consistency_check.py execute here. The 15‑min sweep (POST /consistency/sweep/remote-all) runs this checkout when activity-core triggers via ops-bridge.
railiance01 (92.205.62.239)	activity-core k8s runtime (triage, legacy-meter review, state_hub resolver). Code ships via container image, not host git pull alone.

Evidence (2026-07-09): ~150k weekly legacy GET calls; railiance01 host checkout was 050cbbc; git pull failed with .git/FETCH_HEAD permission denied.

Target commits (minimum):

Repo	Commit	Deliverable
state-hub	14b7d84	/workplans callers, identity headers, workplan-index 410
activity-core	5a92923	/workplans resolver, progress workplan_id only

Out of scope

• Retiring additional /workstreams routes (STATE-WP-0070 T02).
• MCP alias removal (STATE-WP-0070 T03).
• State Hub cluster image rebuild on railiance01 (in-cluster API is separate from workstation sweep path unless explicitly required).

Task: WSL workstation baseline

id: STATE-WP-0071-T01
status: done
priority: high
state_hub_task_id: "760693d4-38df-4949-a739-a0be62f2d6a9"

On the operator WSL machine (State Hub home):

1. bridge status — state-hub-railiance01 must be connected.
2. cd ~/state-hub && git pull --ff-only origin main — verify ≥ 14b7d84.
3. curl -fsS http://127.0.0.1:8000/state/health — start API with make api if down.
4. Record git log -1 --oneline in progress evidence.

Done when workstation state-hub is at target commit and hub API is healthy.

Task: ops-bridge tunnel health

id: STATE-WP-0071-T02
status: done
priority: high
state_hub_task_id: "87e401e1-3b92-4a10-a275-016835be1099"

Restore tunnel mesh required for railiance01 → workstation sweep:

1. bridge status — all production tunnels connected (especially state-hub-railiance01, state-hub-mcp-railiance01).
2. If stale/stopped: bridge up state-hub-railiance01 then bridge check state-hub-railiance01.
3. From WSL: confirm activity-core bridge path reaches hub (actcore-state-hub-bridge pattern per docs/consistency-sweep-runbook.md).

Done when state-hub-railiance01 is connected and hub health is reachable through the bridge URL used by activity-core.

Task: railiance01 git checkout repair and pull

id: STATE-WP-0071-T03
status: done
priority: high
state_hub_task_id: "c86fd968-231d-496b-8866-9d7cae289a99"

SSH to railiance01 (ssh railiance01 / tegwick@92.205.62.239):

1. Fix .git ownership if needed: sudo chown -R "$(whoami):$(id -gn)" ~/state-hub/.git ~/activity-core/.git
2. cd ~/activity-core && git fetch origin main && git pull --ff-only origin main — verify ≥ 5a92923.
3. cd ~/state-hub && git pull --ff-only origin main — verify ≥ 14b7d84 (host checkout; used for operator scripts/docs).
4. Log both SHAs in progress evidence.

Done when both repos pull cleanly and meet minimum commits.

Blocked path: if permission fix fails, flag needs_human with exact error.

Task: activity-core k8s image rollout (railiance01-prod)

id: STATE-WP-0071-T04
status: done
priority: high
state_hub_task_id: "a57a53b6-77b6-4f67-9065-5d451a28b0d5"

Roll resolver/progress changes into the running cluster per activity-core/k8s/railiance/README.md:

1. Build activity-core:railiance01-prod from ~/activity-core at ≥ 5a92923.
2. docker save + scp to railiance01 + k3s ctr images import.
3. rsync k8s manifests if changed; kubectl apply -f k8s/railiance/20-runtime.yaml.
4. Run actcore-sync job; restart actcore-worker, actcore-api, actcore-event-router.
5. kubectl -n activity-core get pods — all runtime pods ready.

Done when actcore-api health check passes and worker pods run the new image digest.

Task: Consistency sweep canary and legacy-meter capture

id: STATE-WP-0071-T05
status: done
priority: high
state_hub_task_id: "61f5a0b7-e382-455f-a06b-7449573d833b"

Verify end-to-end caller migration:

1. From WSL, trigger one sweep: curl -s -X POST http://127.0.0.1:8000/consistency/sweep/remote-all \ -H "Content-Type: application/json" -d '{"max_seconds": 300}'
2. Confirm progress event consistency_sweep_remote_all with exit_code: 0 and no automation_error.
3. Capture evidence: python ~/state-hub/scripts/capture_legacy_meter_evidence.py --days 7 → docs/evidence/legacy-meter-weekly-review-YYYYMMDD.json
4. Success signals:
   • state-hub.fix-consistency appears in component buckets (not only unknown).
   • GET /workstreams/* 7-day window counts trend down vs 2026-07-08 baseline.

Done when evidence file is committed and shows attribution + declining legacy REST.

Task: Closeout — update parent workplan and ops message

id: STATE-WP-0071-T06
status: done
priority: medium
state_hub_task_id: "a1689d68-6000-4d48-8c2f-e3412384244d"

1. Mark STATE-WP-0070 T01 done if success criteria met.
2. Reply on ops-warden thread 467153f3-30ee-44c0-b89c-7f99fdf9f005 with deploy result (SHAs, sweep outcome, meter snapshot path).
3. Update docs/evidence/railiance01-state-hub-deploy-20260709.md checklist.
4. Log hub progress milestone; move this workplan to finished.

Done when STATE-WP-0071 is finished and STATE-WP-0070 T01 reflects deploy complete.

Sequencing

T01 WSL baseline ──┐
T02 bridge health ─┼──► T03 railiance01 git ──► T04 k8s rollout ──► T05 meter capture ──► T06 closeout

T01 and T02 can run in parallel. T04 depends on T03. T05 depends on T04 and T01.

Acceptance criteria

• Workstation and railiance01 checkouts at target commits.
• activity-core k8s runtime healthy on railiance01.
• Legacy-meter 7-day capture shows state-hub.fix-consistency attribution.
• GET /workstreams/* weekly counts materially below 2026-07-08 baseline (exact threshold: operator judges ≥50% drop on top three GET keys, or zero after one full week post-deploy).

Automatic implementation notes (Ralph / agent)

• Use warden sign / SSH cert if raw ssh railiance01 fails; do not request secrets via State Hub messages.
• Prefer bridge MCP or CLI for tunnel ops.
• Human gate: sudo chown on railiance01 if non-interactive SSH cannot elevate.
• Do not retire additional legacy routes in this workplan (STATE-WP-0070 T02).
• After workplan file is approved, run statehub fix-consistency --repo state-hub to register tasks in the hub DB.