Apache Kafka Practice Test

What the Apache Kafka is and how it works

According to Confluent, Inc., “Question types vary, and include multiple-choice, matching, list order.”^[1]

Confluent, Inc. runs the Apache Kafka certification program. The CCAAK exam is aimed at people who administer Kafka clusters, and the CCDAK exam is aimed at developers who build applications on Kafka. Each exam costs USD 150 and has a time limit of 90 minutes.

Exams are taken online and remotely proctored. You need a webcam, a microphone and Google Chrome, and you can test from anywhere that meets Confluent's internet, security and privacy requirements. Question types vary and include multiple-choice, matching and list-order. Results appear immediately on the testing screen.

No training class is required, although Confluent says recommended training helps with preparation. The account name must match your government ID. If you need to move your appointment, reschedule or cancel at least 5 calendar days ahead, because rescheduling is not available within 5 days of the appointment date. If you do not pass, you must wait 7 days before you can buy and take another exam, and you repurchase it in your training account.

The certification is valid for 2 years, and you re-certify every 2 years. Passing scores and question counts are not in the facts we have verified, so check the official Confluent certification page for current details before you book.

Apache Kafka Exam at a Glance

Apache Kafka content outline and where to practise each area

The content areas below are copied from the official Apache Kafka content outline published by Confluent, Inc., with the share of the exam each area carries. The right-hand column shows which of our practice sets covers it.

Worked examples: how to reason through Apache Kafka questions

These are practice questions written for this site, not questions from the real exam. Each explanation shows why the right answer is right and why the others are wrong.

Example 1

Question: What is the function of the Kafka `__transaction_state` internal topic?

1. Stores the state of ongoing and completed transactions managed by transaction coordinators
2. Logs all broker configuration changes
3. Tracks consumer group rebalance events
4. Stores compacted key offsets for cleanup

Answer: A. Stores the state of ongoing and completed transactions managed by transaction coordinators

Explanation:

The correct answer is that __transaction_state stores the state of ongoing and completed transactions managed by transaction coordinators. Kafka keeps this in an internal topic so a coordinator can recover after a failure and still know which transactions were open.

The broker configuration option is wrong because configuration changes are not logged in this topic. The rebalance option is wrong because consumer group coordination is tracked separately from transactions. The compacted key offsets option is wrong because it describes log cleanup state, which is not what a transaction coordinator needs.

Rule to remember: when a feature needs durable state that survives a restart, Kafka usually stores it in an internal topic. Match the topic name to the feature it supports.

Example 2

Question: What does a 'stream-table join' in Kafka Streams do?

1. Enriches each stream record with the current table value for the matching key
2. Merges two streams by timestamp
3. Aggregates stream records into table format
4. Joins two tables producing a new table

Answer: A. Enriches each stream record with the current table value for the matching key

Explanation:

A stream-table join enriches each stream record with the current table value for the matching key. The table side holds the latest value per key, so each incoming record triggers a lookup and the result carries both pieces of data.

Merging two streams by timestamp is wrong because that describes combining streams, not a lookup against a table. Aggregating stream records into table format is wrong because that is an aggregation, which builds a table rather than joining against one. Joining two tables to produce a new table is wrong because that is a table-table join.

Rule to remember: name the two inputs first. A stream and a table gives an enriched stream, because the stream record drives the lookup.

Example 3

Question: In Kafka, what is referred to as a broker?

1. Zookeeper
2. Message
3. Subscriber
4. Server

Answer: D. Server

Explanation:

In Kafka, a broker is a server. A Kafka cluster is made up of broker servers that receive messages from producers, store them, and serve them to consumers. That is why the correct answer is Server.

Zookeeper is wrong because it is a separate coordination service, not the thing that stores and serves messages. Message is wrong because a message is the unit of data a broker handles, not the handler. Subscriber is wrong because it describes the consuming side, which in Kafka terms is a consumer, and a consumer reads from the broker rather than being one.

Rule to remember: producers write, consumers read, and brokers sit between them as the servers that hold the data.

How to use these practice tests

Start with the official Confluent outline for the exam you are taking, and mark each content area as comfortable, shaky or unknown. Spend your first sessions on the shaky and unknown areas rather than rereading what you already know.

Then work through our sets by area. Use Apache Kafka Test for fundamentals and application development, Apache Kafka Kafka Streams for streams and tables, and Apache Kafka Kafka Connect for connectors. After each question, read the explanation even when you answered correctly, and write down the rule behind the right option.

When you miss a question, go back to the Confluent documentation or recommended training for that topic instead of just memorising the answer. Our questions are written by us and are not real exam questions, so the goal is understanding, not recall.

Close to test day, check your webcam, microphone and Google Chrome, and make sure your account name matches your government ID. Do a final timed run through the sets so the 90 minutes feel familiar.

Sources & last verified

1. Confluent, Inc. — number of questions, retake policy, validity, renewal (official source). “Question types vary, and include multiple-choice, matching, list order.” Verified Oct 22, 2026.
2. Confluent, Inc. — time limit, exam fee (official source). “Duration: 90 minutes” Verified Oct 22, 2026.
3. Confluent, Inc. — time limit, exam fee (official source). “Duration: 90 minutes” Verified Oct 22, 2026.
4. Confluent, Inc. — other (official source). “apache kafka application development 28%” Verified Oct 22, 2026.

How we verify: editorial & fact-checking policy. Report an error: contact us.