Confluent · CCAAK
Confluent Certified Administrator for Apache Kafka (CCAAK) Practice Test
Validates expertise in managing Apache Kafka clusters in production, covering Kafka fundamentals, cluster configuration, security, deployment architecture, Kafka Connect administration, observability, and troubleshooting.
Exam Details
Questions
624
Duration
90 minutes
Passing Score
70%
Difficulty
AssociateLast Updated
Feb 2026
Confluent Certified Administrator for Apache Kafka (CCAAK) Practice Exam Preparation
Use this CCAAK practice exam to prepare for Confluent Certified Administrator for Apache Kafka (CCAAK) with realistic questions, detailed explanations, and focused study modes. The practice bank includes 624 questions for Confluent CCAAK, so you can review the exam steadily instead of relying on one long cram session.
As you practice, pay extra attention to patterns in your missed answers. Start with short sessions to identify weak areas, then move into timed quizzes once your accuracy is consistent.
The explanations are especially useful when you want to connect exam wording to the responsibilities and scenarios described in the official certification guidance. Use the free preview first, then unlock the full question bank when you are ready to build a complete study routine.
Exam Domain Breakdown
Apache Kafka Cluster Configuration22%
Apache Kafka Fundamentals15%
Apache Kafka Security15%
Troubleshooting15%
Kafka Connect Administration12%
Deployment Architecture12%
Observability10%
Exam Overview
The Confluent Certified Administrator for Apache Kafka (CCAAK) is a professional certification that validates the skills required to deploy, configure, monitor, and maintain Apache Kafka clusters in production environments. It covers a broad spectrum of administrative competencies, including broker and topic configuration, ZooKeeper management, security implementation (SSL/TLS, SASL, ACLs), Kafka Connect administration, Schema Registry, observability practices, and production troubleshooting. The certification is offered by Confluent, the company founded by the original creators of Apache Kafka, and is recognized across the industry as a credible benchmark for Kafka operations expertise.
The exam tests both conceptual understanding and scenario-based knowledge, requiring candidates to demonstrate proficiency with real-world challenges such as managing consumer group rebalances, diagnosing replication health, configuring listener protocols correctly, and resolving consumer lag. Candidates must understand the roles of brokers, leaders, group coordinators, and how components like Kafka Streams and ksqlDB interact within the broader ecosystem. The certification expires after two years, requiring recertification to remain current.
Who Should Take This Exam
The CCAAK is designed for professionals who are responsible for the day-to-day administration and operation of Apache Kafka clusters. This includes platform engineers, site reliability engineers (SREs), DevOps engineers, and infrastructure administrators who manage Kafka in self-managed, Kubernetes-based, or cloud-hosted environments. Candidates typically have hands-on experience with Kafka CLI tooling and configuration files, and are comfortable diagnosing issues such as replication lag, consumer timeouts, and partition imbalances.
The certification is well-suited for professionals who want to formalize their Kafka administration skills and distinguish themselves in the job market. It is not intended for developers building Kafka-based applications (who would be better served by the CCDAK developer certification), but rather for those responsible for the health, security, and operational performance of Kafka infrastructure.
Prerequisites
Confluent does not enforce formal prerequisites to register for the CCAAK exam. However, candidates are strongly recommended to have practical, hands-on experience running Kafka in a production or production-like environment before attempting the exam. This includes comfort with broker configuration files, CLI tools (kafka-topics, kafka-consumer-groups, kafka-configs, etc.), and experience troubleshooting common operational issues such as under-replicated partitions, consumer lag, and connectivity errors.
Familiarity with Kafka's core components — brokers, producers, consumers, consumer groups, ZooKeeper (or KRaft mode), Kafka Connect, and Schema Registry — is essential. Confluent recommends reviewing the official online study guide and, optionally, completing their formal training courses (available in live and self-paced formats) prior to sitting the exam. Most candidates report studying between 30 and 120 hours depending on their existing Kafka experience.
Exam Format
The CCAAK is a 90-minute, proctored, multiple-choice exam delivered online or at authorized testing centers worldwide. The exam consists of multiple-choice and multi-select questions, with the total question count reported at approximately 40–60 questions depending on the exam version. Remote delivery requires a webcam for proctor monitoring throughout the session. The exam is administered in English only, and results are displayed immediately upon completion.
The passing score is 70%, and the cost per attempt is $150 USD. The certification is valid for two years, after which recertification is required. Upon passing, candidates receive a Confluent digital badge and certificate and are authorized to use the certification title and logo in professional materials.
Skills Measured
- 1.Apache Kafka Cluster Configuration (22%): Covers broker configuration parameters, topic creation and management, partition assignment strategies, replication factor settings, log retention and compaction policies, and producer/consumer configuration tuning including acks, min.insync.replicas, and offset management.
- 2.Apache Kafka Fundamentals (15%): Tests understanding of core Kafka architecture including brokers, topics, partitions, offsets, producers, consumers, consumer groups, leader election, and the role of ZooKeeper or KRaft in cluster coordination.
- 3.Apache Kafka Security (15%): Covers securing Kafka clusters using SSL/TLS for encryption in transit, SASL mechanisms (PLAIN, SCRAM, GSSAPI) for authentication, ACL-based authorization, and ZooKeeper-stored ACL management. Includes listener configuration and security protocol mapping.
- 4.Troubleshooting (15%): Focuses on diagnosing and resolving common Kafka operational issues including under-replicated partitions, consumer lag, ISR shrinkage, offset reset scenarios, rebalance storms, and advertised.listeners misconfigurations.
- 5.Kafka Connect Administration (12%): Covers deploying and managing Kafka Connect in standalone and distributed modes, configuring source and sink connectors, managing connector offsets, understanding worker configuration, and troubleshooting connector failures.
- 6.Deployment Architecture (12%): Addresses cluster topology design, broker placement, network configuration, hardware sizing considerations, multi-datacenter replication (MirrorMaker 2), and deployment on bare metal, VMs, Kubernetes, and managed platforms.
- 7.Observability (10%): Covers monitoring Kafka clusters using JMX metrics, key broker and consumer group metrics to track (under-replicated partitions, request latency, consumer lag), log monitoring, alerting strategies, and integration with monitoring tools.
Study Tips
- Download and thoroughly study the official Confluent CCAAK study guide from the Confluent training portal (training.confluent.io), as it defines the exact domain objectives and weights the exam is built around.
- Practice hands-on with a local Kafka cluster (Docker Compose or Confluent's community packages) to gain fluency with CLI tools like kafka-topics.sh, kafka-consumer-groups.sh, kafka-configs.sh, and kafka-acls.sh — the exam includes scenario-based questions that require operational knowledge.
- Master the interaction between acks, min.insync.replicas, and replication factor — this is a high-frequency topic on the exam and a common source of confusion. Work through failure scenarios manually to understand ISR behavior under different configurations.
- Focus significant study time on security configuration: understand how listeners and advertised.listeners are configured, the difference between SASL mechanisms (PLAIN vs. SCRAM vs. GSSAPI), how SSL keystores and truststores work, and how ACLs are applied to topics, groups, and transactional IDs.
- Use Confluent's free self-paced training courses on Confluent Training (training.confluent.io) as a structured study path — particularly the Apache Kafka Administration course — to cover all exam domains systematically.
- Practice interpreting Kafka metrics (JMX): know which metrics signal under-replication, consumer lag, high request latency, and broker instability. The observability domain tests your ability to diagnose cluster health from metric data.
- When reviewing Kafka Connect, ensure you understand the differences between standalone and distributed mode, how exactly-once semantics work with connectors, and how to troubleshoot a connector in a FAILED state using the REST API.
Career Benefits
Earning the CCAAK demonstrates verified expertise in Apache Kafka administration, a skill set in high demand as organizations across financial services, e-commerce, telecommunications, and technology sectors scale their event streaming infrastructure. Kafka administrators and platform engineers with this credential are well-positioned for roles such as Senior Kafka Administrator, Platform Engineer, Data Infrastructure Engineer, and Site Reliability Engineer. The certification serves as a credible differentiator in both salary negotiations with current employers and job applications with new ones.
Apache Kafka skills consistently command above-average compensation in the data engineering and platform engineering space, with experienced Kafka administrators in North America typically earning $130,000–$180,000+ USD annually. The CCAAK pairs well with cloud certifications (AWS, GCP, Azure) and complements the Confluent Certified Developer for Apache Kafka (CCDAK) for professionals seeking full-stack Kafka expertise. As organizations increasingly adopt event-driven architectures, demand for credentialed Kafka operators continues to grow.
Sample Questions
5 sample questions with answers and explanations. Start a practice session to test yourself across all 624 questions.
Preview — answers shown1. An administrator creates a topic with the command: kafka-topics.sh --bootstrap-server localhost:9092 --create --topic events --partitions 12 --replication-factor 3 --config segment.index.bytes=5242880. What is the purpose of the segment.index.bytes configuration? (Select one!)
AIt determines the maximum message size that can be indexed within a segment
BIt sets the frequency at which offset entries are added to the segment index
CIt controls the maximum size of the offset-to-file-position index file for each log segment
DIt specifies the total size of all index files across all partitions
Explanation
The segment.index.bytes parameter controls the maximum size of the offset index file that maps logical offsets to physical file positions within each log segment. When this size is reached, the index becomes full and a new segment must be created. The frequency of index entries is controlled by index.interval.bytes, not segment.index.bytes. This setting doesn't determine message size limits, which are controlled by max.message.bytes. It applies per segment, not across all partitions.
2. A data engineering team deploys a JDBC source connector in Kafka Connect distributed mode with tasks.max=8. The connector is reading from a single MySQL table with no partitioning column configured. After deployment, only one task is running despite having 8 available workers. What is the most likely cause? (Select one!)
ADistributed mode requires the connector.class to explicitly enable multi-task support
BThe config.storage.topic has insufficient partitions to support 8 concurrent tasks
CThe tasks.max setting requires a corresponding tasks.min setting to take effect
DThe connector can only create one task when reading from a single table without a partitioning column
Explanation
JDBC source connectors can only parallelize reads when a partitioning column is specified in the configuration. Without a partitioning column, the connector cannot split the table into multiple independent tasks and will generate only one task regardless of the tasks.max setting. The connector.class does not require explicit multi-task enablement as task generation is determined by connector logic based on the source data structure. There is no tasks.min configuration parameter in Kafka Connect. The config.storage.topic stores connector configurations and is unrelated to task execution capacity, which is determined by available workers and connector-generated task count.
3. An operations team monitors consumer lag using kafka-consumer-groups.sh --describe --group app-consumers. The output shows: TOPIC=orders, PARTITION=0, CURRENT-OFFSET=10000, LOG-END-OFFSET=15000, LAG=5000, CONSUMER-ID=consumer-1-abc123, HOST=/192.168.1.10. The team also observes that LOG-END-OFFSET increases by 1000 per minute while CURRENT-OFFSET increases by 500 per minute. What action should be taken? (Select one!)
AOptimize the consumer's message processing logic to improve throughput
BAdd more consumer instances to the consumer group to parallelize consumption
CIncrease the consumer's max.poll.records to consume more messages per poll
DIncrease the number of partitions for the orders topic to improve parallelism
Explanation
The consumer is falling further behind with lag increasing by 500 messages per minute (consuming 500/min while producers send 1000/min). Since only one partition is shown and only one consumer is assigned to it, the bottleneck is the consumer's processing speed. Optimizing message processing logic to improve throughput is the most direct solution. Increasing max.poll.records allows fetching more records per poll but does not help if processing is the bottleneck. Adding more consumers to the group will not help because only one partition exists, and only one consumer can be assigned to it. Increasing partitions would enable more consumers to work in parallel, but requires producer changes for key distribution and may not be immediately feasible.
4. A data engineering team configures a Kafka Streams application with processing.guarantee=exactly_once_v2 and num.standby.replicas=1. The application performs stateful aggregations using RocksDB state stores. During a rolling deployment of new application versions, the team observes that state restoration takes 15 minutes per instance as RocksDB restores from changelog topics. Which two approaches reduce state restoration time during deployments? (Select two!)
Multiple correct answersAIncrease num.standby.replicas to 2 to maintain more shadow copies of state stores on other instances
BConfigure state.dir to use SSD storage instead of HDD for faster disk I/O during restoration
CSet cache.max.bytes.buffering to a larger value to reduce changelog topic size and restoration data volume
DEnable RocksDB's write-ahead log (WAL) to checkpoint state incrementally during processing
EUse GlobalKTable instead of KTable for aggregations to eliminate state store restoration requirements
Explanation
Increasing num.standby.replicas creates additional shadow state store copies on different instances. During instance failure or deployment, Kafka Streams can promote a standby replica to active status nearly instantly, avoiding the need to restore state from changelog topics. Using SSD storage for state.dir significantly reduces restoration time because RocksDB state restoration is disk I/O intensive, involving reading large volumes of data from changelog topics and rebuilding local RocksDB databases. Increasing cache.max.bytes.buffering affects record caching before writing to state stores and changelog topics, but does not reduce the total data volume that must be restored from changelogs during recovery. Kafka Streams specifically disables RocksDB's WAL since changelog topics serve this purpose; enabling it would add overhead without improving restoration time. GlobalKTable creates fully replicated state stores on every instance, which actually increases restoration time and resource usage rather than reducing it; GlobalKTables are for reference data, not aggregations.
5. An administrator configures log compaction for a user profile topic with cleanup.policy=compact, log.cleaner.min.cleanable.ratio=0.5, and log.cleaner.delete.retention.ms=86400000. A user deletion event publishes a tombstone record with a null value. After 12 hours, the tombstone is still present. After 36 hours, the tombstone has been removed. Which statement explains this behavior? (Select one!)
ATombstones are retained for log.cleaner.delete.retention.ms duration before becoming eligible for removal during compaction
BTombstones are immediately removed but the consumer lag caused the delayed observation
CThe log.cleaner.min.cleanable.ratio threshold was not reached until after 24 hours elapsed
DThe log.cleaner.threads=1 default caused processing delays in tombstone removal
Explanation
Tombstone records (null values) in compacted topics are retained for log.cleaner.delete.retention.ms duration (86400000 milliseconds = 24 hours) to ensure all consumers have time to observe the deletion. After this retention period expires, the next compaction cycle removes the tombstone. The log.cleaner.min.cleanable.ratio determines when compaction runs but does not affect tombstone retention duration. Consumer lag would not explain why the tombstone disappeared from the topic itself. The number of cleaner threads affects compaction throughput but does not change the configured retention behavior.
$17.99
One-time access to this exam
Full access to all 624 questions
Or $15/mo for all 253 exams
Detailed explanations
Free preview stays available