🚖 Mini-Uber Backend (Microservices Architecture)

A production-grade event-driven microservices project inspired by Uber, built with Spring Boot, Kafka, Eureka, Gateway, Zipkin, Prometheus/Grafana, Redis, and Postgres.

---

🏗️ Architecture

🏗️ Kafka topics map

Services

• API Gateway (Spring Cloud Gateway + JWT Security + Resilience4j)
• Discovery Server (Eureka Service Registry)
• User Service (Postgres + JPA)
• Driver Service (Postgres + JPA)
• Ride Service (Postgres + Kafka producer/consumer)
• Matching Service (Kafka consumer/producer for driver assignment)
• Payment Service (Postgres + Kafka consumer/producer)
• Tracking Service (Redis for live driver tracking)
• Notification Service (Kafka consumer, logs notifications)

Event Pipeline (Kafka)

1. Ride requested → published by ride-service
2. Consumed by matching-service → driver assigned
3. Driver assignment consumed by ride-service → ride started
4. Ride completed → published by ride-service
5. Consumed by payment-service → payment created
6. Consumed by notification-service → notification sent

---

🏗️ Component Diagram

---

🚀 Features

✅ Microservices with service discovery & load balancing
✅ Secure Gateway with JWT Authentication
✅ Kafka event-driven communication between services
✅ Distributed tracing with Zipkin (traceId across HTTP + Kafka)
✅ Metrics & monitoring with Prometheus + Grafana
✅ Resilience4j Circuit Breakers & fallback in API Gateway
✅ Redis caching for driver tracking
✅ PostgreSQL persistence with JPA + Flyway migrations
✅ Centralized logging & observability

---

🛠️ Tech Stack

• Spring Boot 3, Spring Cloud
• Postgres, Redis
• Apache Kafka
• Zipkin, Micrometer, Prometheus, Grafana
• Resilience4j
• Docker (optional, can run locally)
• JUnit/Testcontainers for integration tests

---

🏗️ Deployment and ports

---

🏗️ Sequence Diagram — Ride lifecycle

---

⚡ Run Locally

1. Start Infrastructure

Make sure you have:

• Postgres running on:
  • user-service → 5432
  • driver-service → 5433
  • ride-service → 5434
  • payment-service → 5435
• Redis on localhost:6379
• Kafka broker on localhost:29092
• Zipkin on http://localhost:9411
• Prometheus on http://localhost:9090
• Grafana on http://localhost:3000

These can be started with docker-compose -f infra/docker-compose.yml up.

---

2. Run All Services

One command 🚀:

./run-all.sh

---

3. 🧪 Testing the Workflow

Create User

curl -X POST http://localhost:8080/api/users \
     -H "Content-Type: application/json" \
     -d '{"name":"Alice","email":"alice@mail.com"}'

Request Ride

curl -X POST http://localhost:8080/api/rides \
     -H "Authorization: Bearer <JWT>" \
     -H "Content-Type: application/json" \
     -d '{"userId":1,"startLocation":"A","endLocation":"B"}'

Kafka Event Flow

ride-service → matching-service → ride-service → payment-service → notification-service

Complete Ride

curl -X PATCH http://localhost:8080/api/rides/{id}/complete \
     -H "Authorization: Bearer <JWT>"

🛠️ Troubleshooting

• Kafka UnknownHost → use localhost:29092 (host) vs kafka:9092 (Docker)
• JWT Weak Key → use Keys.secretKeyFor(SignatureAlgorithm.HS256)
• DB Connection Refused → check Postgres container ports + spring.datasource.url
• Prometheus scrape errors → verify /actuator/prometheus endpoint

📸 Screenshots

🟦 Eureka — Service Registry

Shows all services (user, driver, ride, payment, matching, tracking, notification) registered with discovery.

📊 Grafana — Observability Dashboard

Custom dashboards for API latency, Kafka throughput, and JVM metrics.

📈 Prometheus — Metrics Source

Scraping from all microservices via /actuator/prometheus.

🔎 Zipkin — Distributed Tracing

End-to-end trace of a ride request across Gateway → Ride → Matching → Payment → Notification.