Real-Time Order Notification System

Java · Spring Boot · Apache Kafka · MySQL · Docker

⚠️ Zero Lombok — works out of the box in STS / IntelliJ / VS Code

---

Architecture

Client (Postman)
      │  REST POST /api/v1/orders
      ▼
┌─────────────────────┐     publishes      ┌──────────────────────┐
│   Order Service     │ ─────────────────▶ │   Kafka Topic        │
│   (Port 8081)       │                    │   "order-events"     │
│                     │                    │   (3 partitions)     │
│  REST API           │                    └──────────┬───────────┘
│  Kafka Producer     │                               │ consumes
│  MySQL (orderdb)    │                    ┌──────────▼───────────┐
└─────────────────────┘                    │ Notification Service │
                                           │ (Port 8082)          │
                                           │                      │
                                           │ Kafka Consumer       │
                                           │ Idempotency Check    │
                                           │ MySQL (notificationdb│
                                           └──────────────────────┘

---

Tech Stack

Technology	Usage
Java 17	Core language (no Lombok)
Spring Boot 3.2	Application framework
Spring MVC	REST API
Apache Kafka	Event streaming
MySQL 8.0	Persistent storage
Docker Compose	Local infra setup
JUnit + Mockito	Unit testing

---

How to Run

Step 1 — Start Infrastructure

docker-compose up -d

Starts: Kafka + Zookeeper + MySQL + Kafka UI

Step 2 — Run Order Service

cd order-service
mvn spring-boot:run

→ http://localhost:8081

Step 3 — Run Notification Service

cd notification-service
mvn spring-boot:run

→ http://localhost:8082

Step 4 — Test APIs

Import OrderNotificationSystem.postman_collection.json into Postman.

Step 5 — Watch Kafka Messages

Open Kafka UI → http://localhost:8080 → Topics → order-events → Messages

---

REST Endpoints

Method	Endpoint	Description
POST	/api/v1/orders	Place a new order
GET	/api/v1/orders	Get all orders
GET	/api/v1/orders/{orderId}	Get order by ID
GET	/api/v1/orders?email={email}	Get orders by email
PATCH	/api/v1/orders/{orderId}/status	Update order status

Sample Request

POST http://localhost:8081/api/v1/orders
{
  "customerName": "Priya Ghadge",
  "customerEmail": "priya@test.com",
  "productName": "Laptop",
  "quantity": 1,
  "price": 75000.00
}

---

Key Design Decisions

1. Kafka message key = orderId All events for the same order go to the same partition → ordering guaranteed per order.

2. Manual offset acknowledgment Offset committed ONLY after successful processing. If consumer crashes, Kafka redelivers.

3. Idempotency pattern Unique key = orderId_status. Duplicates detected and skipped safely.

4. Constructor injection (no @Autowired) All dependencies injected via constructor — testable and explicit.

---

Interview Talking Points

"Why Kafka over REST between services?" Kafka decouples the services. Order service doesn't wait for notification to complete. If notification service is down, messages queue in Kafka and process when it recovers.

"How do you handle duplicates?" Kafka delivers at-least-once. I store an idempotency key (orderId + status) in DB. Before processing, I check if this key exists — if yes, I skip it safely.

"What if the consumer crashes mid-processing?" Manual ack (AckMode.MANUAL_IMMEDIATE) means offset is only committed after success. Kafka redelivers to another consumer in the group — no data loss.

"How would you scale this?" Increase partitions to match consumer instances. Deploy on Kubernetes with HPA. 3 partitions = 3 parallel consumer threads processing simultaneously.

---

Author

Priya Ghadge | Java Developer | Spring Boot · Kafka · Microservices LinkedIn: https://linkedin.com/in/priya-ghadge