Operate live, recover safely, and prove readiness

This guide explains the enterprise operating layer across HavenSuite: real-time hotel events, offline staff actions, integration health, multi-property control, the secure guest journey, request tracing, scheduler and queue evidence, privacy review, and release verification.

1. Live Operations Command Center

Open Operations to see role-filtered arrivals, room work, maintenance, handovers, finance alerts, intelligence updates, and integration incidents. The page uses server-sent events when available and falls back to safe polling.

1. Review Arrivals and Ready rooms before the shift.
2. Open an event only when its action belongs to your role.
3. Acknowledge a warning when a named staff member owns the next action.
4. Keep manual room, payment, and front-desk workflows available during provider failure.

2. Offline Staff Work

The staff outbox supports selected housekeeping, maintenance, and room-condition updates when connectivity is unstable. It never queues payments, signatures, evidence uploads, or destructive actions.

Status	Meaning	Operator action
Saved offline	Stored only for this device, hotel, and user.	Reconnect without repeating the action.
Conflict	Another user changed the record.	Review the current server record.
Synced	The server accepted the idempotent action.	No further action.

3. Integration Health

Use Administration → Integration Health to check channels, payments, messaging, calendars, webhooks, and Haven Intelligence. States are Healthy, Degraded, Failing, Manual, or Not configured. Health details never expose credentials.

4. Multi-property Portfolio

Hotel Groups compares occupancy, ready rooms, movement, revenue, and exceptions. Revenue stays separated by currency. Group membership does not bypass hotel membership; opening a property requires active access to it.

5. Secure Guest Journey

The portal combines stay progress, room assignment, handover status, billing, payment links, requests, concierge, and history under the existing hashed expiring token. Guest requests enter needs_review and never directly mutate rooms, tasks, rates, folios, or reservations.

6. Reliability Center

Check	Healthy evidence	Failure response
Scheduler	app:heartbeat within three minutes.	Run schedule:run every minute.
Queue	No failed jobs and acceptable depth.	Retry only idempotent work.
Backup	Recent completed encrypted archive.	Run backup and rehearse restore.
Storage	Runtime directories are writable.	Correct hosting ownership.
Privacy	Policies exist and requests have owners.	Complete the audited workflow.

7. Request tracing and language

Every web response includes X-Request-ID. Record it in incident notes. The language selector stores a validated session preference and applies document language/direction; untranslated operational labels remain in English.

8. Release verification

php artisan app:heartbeat
php artisan test --stop-on-failure
php artisan access:audit --fail-on-drift
php artisan app:performance-audit --fail-on-budget
npm run build
npm audit --omit=dev
npm run test:visual
composer validate --strict
composer audit

Restore storage/app/installed after local tests if necessary. Production also requires HTTPS, debug disabled, queue workers, scheduler, SMTP, signed webhooks, private storage, and an off-server backup.