Why topics organize messages in Kafka - Why It Works

Introduction

When you send many messages in a system, you need a way to keep them organized. Kafka uses topics to group related messages together so you can find and use them easily.

When you want to separate messages by type, like orders and payments, so they don't mix up.

When you need to let different parts of your app listen only to messages they care about.

When you want to keep messages in order for a specific category, like all messages about one user.

When you want to scale your system by splitting messages into smaller groups.

When you want to store messages for a while and let many apps read them independently.

Commands

This command creates a new topic named 'orders' with 3 partitions. Partitions help organize messages inside the topic and allow parallel processing.

Terminal

kafka-topics --create --topic orders --bootstrap-server localhost:9092 --partitions 3 --replication-factor 1

Expected Output

Created topic orders.

→

--topic - Name of the topic to create

→

--partitions - Number of partitions to split the topic into

→

--replication-factor - Number of copies of the data for fault tolerance

This command lists all topics in the Kafka server so you can see the topics available to organize messages.

Terminal

kafka-topics --list --bootstrap-server localhost:9092

Expected Output

orders

→

--list - List all topics on the server

This command starts a producer that sends messages to the 'orders' topic. Messages sent here are organized under 'orders'.

Terminal

kafka-console-producer --topic orders --bootstrap-server localhost:9092

Expected Output

No output (command runs silently)

→

--topic - Specify the topic to send messages to

This command reads messages from the 'orders' topic starting from the oldest message. It shows how messages are grouped and retrieved by topic.

Terminal

kafka-console-consumer --topic orders --bootstrap-server localhost:9092 --from-beginning --max-messages 3

Expected Output

order1

order2

order3

→

--from-beginning - Read messages from the start of the topic

→

--max-messages - Limit the number of messages to read

Key Concept

Topics in Kafka group related messages so you can organize, find, and process them easily.

Common Mistakes

Trying to send messages without creating a topic first

Kafka requires topics to exist before sending messages; otherwise, the send fails.

Always create the topic with kafka-topics before producing messages.

Using too few partitions for a high volume topic

Too few partitions limit parallel processing and can slow down message handling.

Create topics with enough partitions to handle your expected load.

Confusing topics with partitions

Partitions are parts inside a topic, not separate topics; mixing them up causes wrong assumptions about message grouping.

Remember that topics group messages broadly, partitions split them inside the topic for scaling.

Summary

Create topics to group related messages for better organization.

Use partitions inside topics to allow parallel processing and scaling.

Produce and consume messages by specifying the topic name to keep data organized.

Practice

1. Why does Kafka use topics to organize messages?

easy

A. To group related messages together for easier management

B. To store messages permanently on the producer side

C. To encrypt messages automatically

D. To limit the number of messages sent

Solution

1. Step 1: Understand the purpose of topics in Kafka

   Topics are used to group messages that belong to the same category or stream.

2. Step 2: Identify the role of topics in message management

   Grouping messages helps producers and consumers handle data efficiently and logically.

3. Final Answer:

   To group related messages together for easier management -> Option A

4. Quick Check:

   Topics group messages = A [OK]

Hint: Topics group messages logically for easy handling [OK]

Common Mistakes:

• Thinking topics store messages on producer side
• Confusing topics with encryption features
• Believing topics limit message quantity

2. Which of the following is the correct way to create a topic in Kafka using the command line?

easy

A. kafka topic create myTopic --partitions 3

B. kafka-topics --create --topic myTopic --partitions 3 --replication-factor 1 --bootstrap-server localhost:9092

C. create kafka topic myTopic partitions=3

D. kafka-create-topic --name myTopic --partitions 3

Solution

1. Step 1: Recall the Kafka CLI syntax for topic creation

   The correct command uses 'kafka-topics' with '--create' and required flags.

2. Step 2: Match the command with the correct options

   kafka-topics --create --topic myTopic --partitions 3 --replication-factor 1 --bootstrap-server localhost:9092 matches the official syntax including bootstrap-server, partitions, and replication-factor.

3. Final Answer:

   kafka-topics --create --topic myTopic --partitions 3 --replication-factor 1 --bootstrap-server localhost:9092 -> Option B

4. Quick Check:

   Correct Kafka topic creation syntax = D [OK]

Hint: Use kafka-topics with --create and required flags [OK]

Common Mistakes:

• Using incorrect command names like kafka-create-topic
• Omitting bootstrap-server flag
• Wrong order or missing required flags

3. Given a Kafka topic with 3 partitions, what happens when a producer sends 6 messages without specifying a key?

medium

A. Messages are sent randomly to any partition

B. All messages go to the first partition

C. Messages are distributed evenly across the 3 partitions

D. Messages are rejected because no key is provided

Solution

1. Step 1: Understand default partitioning behavior without a key

   When no key is provided, Kafka uses round-robin to distribute messages evenly across partitions.

2. Step 2: Apply this to 6 messages and 3 partitions

   Messages will be sent 2 per partition, balancing the load.

3. Final Answer:

   Messages are distributed evenly across the 3 partitions -> Option C

4. Quick Check:

   No key means round-robin distribution = A [OK]

Hint: No key means even round-robin distribution [OK]

Common Mistakes:

• Assuming all messages go to one partition
• Thinking messages are rejected without a key
• Believing distribution is random without pattern

4. You have a Kafka topic with 2 partitions but your consumer only reads from partition 0. What is the likely problem?

medium

A. Kafka does not support multiple partitions

B. The topic only has one partition

C. The producer is sending messages only to partition 1

D. The consumer is not configured to subscribe to all partitions

Solution

1. Step 1: Analyze consumer subscription behavior

   If a consumer reads only from partition 0, it likely subscribed only to that partition explicitly.

2. Step 2: Check topic partition count and producer behavior

   The topic has 2 partitions, so consumer should subscribe to both to read all messages.

3. Final Answer:

   The consumer is not configured to subscribe to all partitions -> Option D

4. Quick Check:

   Consumer subscription controls partitions read = C [OK]

Hint: Consumer must subscribe to all partitions to read all messages [OK]

Common Mistakes:

• Assuming topic has only one partition
• Believing Kafka does not support multiple partitions
• Thinking producer controls consumer reading partitions

5. How can partitioning in Kafka topics improve performance when handling large message streams?

hard

A. By allowing parallel processing of messages across multiple consumers

B. By encrypting messages faster

C. By reducing the total number of messages sent

D. By storing messages only on the producer side

Solution

1. Step 1: Understand the role of partitions in Kafka topics

   Partitions split a topic into multiple parts that can be processed independently.

2. Step 2: Connect partitions to performance benefits

   Multiple consumers can read from different partitions in parallel, increasing throughput and speed.

3. Final Answer:

   By allowing parallel processing of messages across multiple consumers -> Option A

4. Quick Check:

   Partitions enable parallelism = B [OK]

Hint: Partitions enable parallel consumer processing [OK]

Common Mistakes:

• Confusing partitioning with encryption
• Thinking partitioning reduces message count
• Believing messages are stored only on producer