DEV Community

Er. Bhupendra
Er. Bhupendra

Posted on

KAFKA IN SPRING BOOT

#beginners #interview #springboot #tutorial

Great Bhupendra 👍
Since you like real understanding (not ratta), I’ll make Kafka super simple + interview ready + practical mindset.

🚀 Apache Kafka – Short & Easy Way to Understand

🧠 First Understand in 1 Line

👉 Kafka = Real-time data post office
It delivers messages between applications very fast & reliably.

🍽️ Simple Example (Udupi Hotel)

Imagine this flow:

Customer App  →  Kafka  →  Udupi Hotel App
Enter fullscreen mode Exit fullscreen mode
  • Customer = Producer
  • Udupi Hotel = Consumer
  • Kafka = Message Broker
  • Topic = Message box

🔁 Real Flow

1️⃣ Customer Places Order 

Customer App ---> [order_topic] ---> Hotel App
Enter fullscreen mode Exit fullscreen mode

Customer = Producer
Hotel = Consumer

2️⃣ Hotel Accepts Order 

Hotel App ---> [order_status_topic] ---> Customer App
Enter fullscreen mode Exit fullscreen mode

Now:
Hotel = Producer
Customer = Consumer

👉 Roles can change!

🏗️ Kafka Architecture (Very Easy) 

Producer → Topic → Consumer
Enter fullscreen mode Exit fullscreen mode

Important Components:

Component Meaning (Simple Language)
Producer Who sends message
Consumer Who receives message
Topic Message container
Broker Kafka server
Zookeeper Manages Kafka (older versions)

🔥 Real Interview-Level Understanding

Why Not Just Use Database?

Use Case DB Kafka
Store order ✅ Yes ❌ No
Real-time notification ❌ No ✅ Yes
Heavy traffic ❌ Slow ✅ Very Fast
Event streaming ❌ Not ideal ✅ Best

👉 Kafka is NOT database replacement
👉 Kafka is event streaming system

🎯 Where Kafka Is Used in Industry

  • Swiggy / Zomato → Orders
  • Amazon → Order tracking
  • Banking → Transactions
  • Uber → Location tracking
  • Microservices communication

Since you're learning microservices 🔥
👉 Kafka is used for asynchronous communication

⚡ Important Kafka Concepts (Must Know)

1️⃣ Topic

Logical container for messages.

Example:

  • order_topic
  • payment_topic

2️⃣ Partition (Very Important 🚨)

Topic is divided into partitions.

Why?
👉 Parallel processing
👉 High scalability

order_topic
   ├── Partition 1
   ├── Partition 2
   └── Partition 3
Enter fullscreen mode Exit fullscreen mode

More partitions = More performance

3️⃣ Offset

Each message has unique number inside partition.

Message 1 → Offset 0
Message 2 → Offset 1
Message 3 → Offset 2
Enter fullscreen mode Exit fullscreen mode

Consumer remembers offset
So no duplicate processing.

4️⃣ Consumer Group

Multiple consumers can read same topic.

order_topic
   ↓
Consumer Group A
   ├── Consumer 1
   └── Consumer 2
Enter fullscreen mode Exit fullscreen mode

Kafka distributes partitions among them.

🧩 Modern Kafka Architecture

⚠️ Important:

👉 New Kafka versions (3.x+) do not require Zookeeper
They use KRaft mode

So in interviews:
If someone says Zookeeper mandatory → ❌ Old knowledge.

🛠️ Setup (Simplified for Understanding)

You basically:

  1. Start Zookeeper (old versions)
  2. Start Kafka
  3. Create Topic
  4. Start Producer
  5. Start Consumer

💡 Very Important Interview Questions

Q1: What is Kafka?

Distributed event streaming platform.

Q2: Why Kafka is fast?

  • Sequential disk write
  • Partitioning
  • Zero copy transfer

Q3: Difference Between Kafka and RabbitMQ?

Kafka RabbitMQ
Event streaming Message queue
High throughput Low latency
Log-based Queue-based
Scalable Less scalable

Q4: What is Exactly Once?

Kafka ensures:

  • No duplicate
  • No message loss

(using idempotent producer + transactions)

🧠 Real Microservices Example (Important for You)

Suppose:

Order Service
Payment Service
Notification Service

Instead of direct REST calls:

Order Service → Kafka → Payment Service
Payment Service → Kafka → Notification Service
Enter fullscreen mode Exit fullscreen mode

Benefits:

  • Loose coupling
  • Scalable
  • Fault tolerant

🎯 When To Use Kafka?

Use Kafka When:

✔ High traffic
✔ Real-time updates
✔ Microservices
✔ Event driven architecture
✔ Log streaming

Do NOT use Kafka when:

❌ Small CRUD app
❌ Simple project

🧾 Final 30-Second Revision

Kafka = Distributed event streaming platform

Main parts:

  • Producer
  • Consumer
  • Topic
  • Partition
  • Offset
  • Broker

Use for:

  • Real-time
  • High traffic
  • Microservices

==========================

================================
Perfect Bhupendra 🔥
Since you’re learning Microservices, this is VERY important for interviews.

🚀 Kafka vs REST in Microservices (Easy + Practical)

🧠 First Understand Core Difference

REST Kafka
Synchronous Asynchronous
Direct call Message-based
Tight coupling Loose coupling
Request/Response Event-driven

📞 1️⃣ REST in Microservices

Flow: 

Order Service  →  Payment Service  →  Notification Service
Enter fullscreen mode Exit fullscreen mode
  • Order calls Payment via HTTP
  • Payment returns response
  • Then Order calls Notification

🔴 Problem:

  • If Payment service is DOWN ❌ → Order fails
  • Services tightly connected
  • Hard to scale under heavy load

🔥 2️⃣ Kafka in Microservices

Flow: 

Order Service → [Kafka: order_created] → Payment Service
Payment Service → [Kafka: payment_done] → Notification Service
Enter fullscreen mode Exit fullscreen mode
  • Order publishes event
  • Payment consumes event
  • No direct dependency

🟢 Benefit:

  • If Payment is down → Kafka stores message
  • When Payment comes back → It processes messages
  • No data loss

🏗️ Real Architecture Comparison

🔴 REST Based Architecture 

Client
  ↓
Order Service
  ↓ HTTP
Payment Service
  ↓ HTTP
Notification Service
Enter fullscreen mode Exit fullscreen mode

👉 Chain dependency
👉 Slow under heavy traffic

🟢 Kafka Based Architecture 

Client
  ↓
Order Service
  ↓
[Kafka]
  ↓
Payment Service
  ↓
[Kafka]
  ↓
Notification Service
Enter fullscreen mode Exit fullscreen mode

👉 Event Driven
👉 Scalable
👉 Fault Tolerant

🎯 Real Interview Question

Q: When should we use REST instead of Kafka?

Use REST when:

✔ Immediate response needed
✔ Simple CRUD
✔ Less traffic
✔ Small system

Example:

  • Login API
  • Fetch user profile

Q: When should we use Kafka?

Use Kafka when:

✔ High traffic
✔ Event driven
✔ Decoupled microservices
✔ Retry mechanism required
✔ Real-time analytics

Example:

  • Order system
  • Payment processing
  • Ride booking
  • Banking transactions

⚡ Important Microservices Concept

🔹 Synchronous vs Asynchronous

REST = Synchronous

Client waits for response.

Kafka = Asynchronous

Producer doesn’t wait.
Message goes to topic.

🧠 Real Industry Example

Amazon Order Flow

  1. Order Created → Kafka
  2. Inventory Updated → Kafka
  3. Payment Processed → Kafka
  4. Email Sent → Kafka

Each service works independently.

💥 Performance Difference

Feature REST Kafka
Scalability Medium Very High
Fault tolerance Low High
Retry Manual Built-in
Throughput Limited Millions/sec
Coupling Tight Loose

🧩 What Big Companies Use?

  • Netflix → Kafka
  • Uber → Kafka
  • LinkedIn → Kafka (created by LinkedIn)
  • Amazon → Event-driven architecture

🎯 Final Easy Rule (Remember This)

If service must WAIT → Use REST
If service must NOT WAIT → Use Kafka

🔥 For Your Interview (Spring Boot Example Answer)

If interviewer asks:

"How would you design order microservice?"

You say:

I would use REST for synchronous calls like login or fetching user data.
For order processing, payment, and notification, I would use Kafka for asynchronous communication to ensure loose coupling and scalability.

🔥 That’s senior-level answer.

==================

==========================

Top comments (0)