Disruption Simulation & Training

Mass flight disruptions are high-stress environments. The Disruption Simulator allows airline platform administrators and operations supervisors to execute end-to-end dry-run scenarios in an isolated sandbox.

Supervisors can simulate hub closures, weather delays, and system out-of-services to:

• Train Agents: Onboard new operators and practice coordination under high concurrency.
• Stress-Test Workflows: Verify how team communication, dispatch speeds, and queues hold up under heavy load.
• Validate Entitlement Policies: Confirm that policy rules (e.g., booking specific hotel tiers for premium passengers) generate correct outcomes before activating them in production.

---

Running a Disruption Simulation

Simulations are executed inside a designated Sandbox Tenant Environment, which mirrors production workflows but directs all partner requests (Amadeus, Hotelbeds, Pomelo) to mock adapters. No real hotel bookings are made, and no real cards are issued.

Step 1: Open the Simulation Dashboard

1. Log in to the platform console using credentials with ADMIN or OPS_SUPERVISOR permissions.
2. Select Operations -> Simulator Dashboard from the navigation sidebar.

Step 2: Configure the Simulation Parameters

Define the scope of the mock disruption:

Field	Description	Example
Target Airport	The IATA code of the hub experiencing the disruption.	SCL (Santiago)
Disruption Type	The root cause (triggers specific automatic policy behaviors).	WEATHER / STRIKE / ATC
Affected Flights	The number of mock flights to delay or cancel.	5 Flights
Passenger Density	Generates varying loads of passengers (affects concurrency).	Normal / High (Swarm Mode)

Step 3: Execute the Simulation

1. Click Start Simulation.
2. The simulator creates a new sandboxed urn:simulation context and auto-generates:
   • Mock flight statuses showing cancellations.
   • Mock passenger manifests containing realistic loyalty tiers, cabin classes, and family groups.
3. The platform's automated detection engine intercepts these mock disruptions and creates corresponding cases in the Sandbox dashboard, notifying active agents.

Step 4: Observe Operator Performance

During the active simulation, supervisors can monitor the live dashboard to track operational efficiency:

• Average Triage Time: The time taken to resolve a case from PENDING to RESOLVED.
• Swarm Lock Effectiveness: Verifying that agents are coordinating passenger lockouts successfully without collisions.
• Outbox Processing Speeds: Observing how the asynchronous queue absorbs transaction surges to partners without exceeding API limits.

---

Simulating Partner Failures (Resilience Drill)

To train operators for worst-case scenarios, the simulator allows you to trigger mock vendor failures:

1. Under Failure Scenarios, check the box for "Simulate Hotel Partner Outage (503)" or "Simulate Card Issuer Outage (500)".
2. Run the simulation.
3. Observe how the platform:
   • Automatically opens circuit breakers.
   • Routes failed attempts to the Manual Review Queue.
   • Initiates compensation sagas (releasing partial bookings) when failures are permanent.
4. Instruct agents to resolve these cases using the Manual Reconciliation Playbook.

---

Concluding a Simulation

1. Once the training session is complete, click End Simulation in the control panel.
2. The system purges all mock data from the sandboxed tenant space.
3. A summary report is generated, displaying agent speeds, policy compliance rates, and queue statistics.