A modular, scalable backend system for searching flights, locking seats, creating bookings, processing payments, applying discounts, generating tickets/invoices, and sending notifications.

Built using TypeScript, Node.js, PostgreSQL, and Redis, with clean class-based architecture, secure JWT auth, Redis caching, and containerized using Docker Compose.

This is a complete backend for a flight booking platform (like MakeMyTrip). It includes 9 decoupled microservices handling:

• Authentication
• Flight search (supports 2-hop logic)
• Seat inventory & locking
• Booking system
• Payment gateway (mock)
• Discount validation
• Ticket & Invoice generation
• Notification dispatch (mock)
• Centralized API Gateway

• Node.js >= 18
• Docker and Docker Compose
• .env files (see below)

NOTE: I have added .env files in the code so that its easy for anyone testing and refrencing the code. DO NOT DO THIS IN PRODUCTION ENVIRONMENT

This starts all 9 services, Postgres DBs, and Redis via Docker Compose.

Run this for running each service individually

Run this for running all services at once

USE POSTMAN COLLECTION FOR TESTING

Each service contains a .env file. Most include:

NOTE: I have added .env files in the code so that its easy for anyone testing and refrencing the code. DO NOT DO THIS IN PRODUCTION ENVIRONMENT

The api-gateway also includes:

Below is a complete end-to-end simulation of a flight booking using Postman or cURL. All APIs are routed via the API Gateway (http://localhost:8080):

1. Signup

   POST /api/auth/signup
Content-Type: application/json

{
"name": "John",
"email": "john@example.com",
"password": "test123"
}


2. Login

   POST /api/auth/login
Content-Type: application/json

{
"email": "john@example.com",
"password": "test123"
}


   Returns a JWT token and userId. Use the token in all further requests as: Authorization: Bearer

3. Search Flights

   GET /api/flights/search?source=DEL&destination=BOM&date=2025-08-05
Authorization: Bearer


4. Check Available Seats

   GET /api/seats/1
Authorization: Bearer


5. Lock a Seat

   POST /api/seats/lock
Authorization: Bearer
Content-Type: application/json

{
"flightId": 1,
"userId": 1,
"seatNumber": "A1"
}


6. Create Booking

   POST /api/bookings/create
Authorization: Bearer
Content-Type: application/json

{
"userId": 1,
"flightId": 1,
"price": 5000,
"seats": 2
}


7. Get Booking Details

   GET /api/bookings/1/1
Authorization: Bearer


8. Initiate Payment

   POST /api/payments/pay
Authorization: Bearer
Content-Type: application/json

{
"bookingId": 1,
"amount": 5000,
"userId": 1
}


9. Apply Discount Code (Optional)

   POST /api/discounts/apply
Authorization: Bearer
Content-Type: application/json

{
"bookingId": 1,
"code": "FLY100"
}


10. Generate Ticket

11. Get Ticket

12. Send Notification

• Flight search results are cached in Redis

• Cache is invalidated if:

  1. Flight has already departed
  2. Flight is in the cancelled list
  3. Flight is fully booked

• Every service (except signup/login) checks for valid JWT

• Booking flow is strictly linear:

  1. Search - Lock - Booking - Pay - Ticket
  2. Discount is optional, all else is mandatory

• Redis ensures race-free seat locking

• Reverse proxy: NGINX or Traefik in front of API Gateway
• Use HTTPS (TLS termination at gateway)
• Run Redis/Postgres as managed services (e.g., GCP Cloud SQL, Memorystore)
• Add retry logic and exponential backoff for network calls
• Add monitoring (Prometheus + Grafana)
• Structured logging (e.g., Winston + FluentBit)