DashBoard/docs/migration/portal-shell-route-view-integration/kill-switch-operations.md

Kill-Switch Operations: Shell Route-View Migration

Last updated: 2026-02-11

Purpose

Provide a rapid, operator-safe mechanism to recover service usability when severe regressions occur after shell cutover.

Trigger Conditions

• Critical route failures on shell core paths (/portal-shell, /api/portal/navigation).
• Multiple P0 smoke failures across Wave A/Wave B pages.
• Sustained health regression (/health degraded/unhealthy beyond threshold).

Kill-Switch Command

• Set PORTAL_SPA_ENABLED=false in deployment environment.
• Restart application workers.

Verification Checklist (must complete in order)

1. GET / responds and routes to legacy portal.
2. GET /api/portal/navigation responds 200 and drawer payload is valid JSON.
3. GET /health reports no new critical errors after rollback.
4. Critical page routes remain reachable: /wip-overview, /resource, /qc-gate, /job-query, /excel-query, /query-tool, /tmtt-defect.

Page-level Partial Kill-Switch

• If issue is route-scoped, patch affected route contract to fallback strategy and redeploy frontend shell assets only.
• Keep unaffected routes in native mode to avoid global disruption.

Escalation

• If kill-switch does not restore stable behavior within 15 minutes, escalate to full rollback runbook and incident bridge.