CCDAK Confluent Certified Developer for Apache Kafka Certification Examination Questions and Answers

The onCompletion() method of Kafka’sCallback interfaceis executedonce the broker acknowledgesthe message. This includes success or failure, and it isinvoked asynchronouslyby the producer.

FromKafka Java Client API Documentation:

“The onCompletion method will be called when the record sent to the server has been acknowledged, or when an error occurs.”

This ensures thatthe record was sent and acknowledged, not just added to a batch or local buffer.

ForSASL/PLAIN authentication over SSL, the correct combination is:

security.protocol=SASL_SSL

sasl.jaas.config=...PlainLoginModule required username=... password=...

FromKafka Security Documentation:

“For SASL/PLAIN over SSL, configure security.protocol=SASL_SSL and provide login module configuration using sasl.jaas.config.”

Options B and D are incorrect because they do not pair the correct protocol with the authentication method.

TheStickyAssignortries to minimize partition movement bypreserving existing assignmentsas much as possible while still achieving a balanced assignment. This improvesconsumer stabilityand reduces rebalances.

From theKafka Consumer Assignor Documentation:

“The StickyAssignor attempts topreserve as many existing assignments as possible, which helps minimize partition movement between rebalances.”

RoundRobinAssignor focuses on even distribution, not stability.

RangeAssignor groups partitions by topic and assigns them consecutively, but can lead to imbalances.

PartitionAssignor is an abstract base class, not an assignor used directly.

Comprehensive and Detailed Explanation From Exact Extract:

✅C. Consumers can consume both committed and uncommitted transactions.By default,Kafka consumers only read committed messagesif they are configured with isolation.level=read_committed. However, if configured as read_uncommitted, theycan also consume uncommitted (potentially aborted) transactional messages.

From Kafka Documentation:

"The isolation.level setting controls whether the consumer will read only committed messages or all messages, including uncommitted messages from ongoing or aborted transactions."

✅D. Information about producers and their transactions is stored in the _transaction_state topic.Kafka uses an internal topic named__transaction_stateto maintain metadata about producer transactions. This topic is essential for tracking thetransaction lifecycle, fencing, and recovery.

From Kafka Internals:

“Kafka stores the state of active and completed transactions in an internal topic called __transaction_state.”

Thebatch.sizeconfig sets themaximum number of bytes to batch per partitionbefore sending. This allows Kafka producers toamortize I/Oand improve throughput.

FromKafka Producer Configuration Docs:

“batch.size is the maximum amount of data per partition the producer will batch before sending.”

buffer.memory sets total memory for the producer, not per-batch.

send.buffer.bytes is aTCP socket buffer, not a Kafka config.

max.block.ms controls blocking time, not size.

To maintainmessage ordering per bank account, themessage key must be the account number. Kafka guaranteesordering per partition, and the partition is selected based on thehash of the key.

FromKafka Producer Docs:

“All records with the same key will be sent to the same partition, and Kafka preserves the order of records in a partition.”

Using account number ensures:

All messages for the same account go to the same partition

Order is maintained

Workload is evenly distributed if account keys are varied

Other options break ordering or introduce unnecessary randomness.

Kafka Streams usestransactional producersto guaranteeexactly-once semantics (EOS). This ensures that both theoutput recordsandstate store updatesare committed atomically, avoiding duplication or partial writes.

FromKafka Streams Documentation > Processing Guarantees:

“Kafka Streams leveragesKafka’s transactional APIsto commit the output records and internal state updatesas a single atomic unit, thereby providing exactly-once semantics.”

Option A is incorrect because log compaction is not disabled for EOS.

Option C incorrectly describes a checkpointing system Kafka Streams does not use.

Option D refers to deduplication, which is not how EOS is achieved in Streams.

RetriableException is a subclass of KafkaException, indicating transient issues thatmight succeed if retried. Examples include:

LeaderNotAvailableException: Happens when metadata is not yet propagated or leader election is in progress.

NotEnoughReplicasException: Happens when the number of in-sync replicas is insufficient. This can be transient if replicas come back online.

From theApache Kafka Java Client documentation:

"These exceptions (like LeaderNotAvailableException, NotEnoughReplicasException) are transient and the client will retry them automatically depending on retry configuration."

RecordTooLargeExceptionandAuthorizationExceptionarenon-retriableas they indicate client-side or permission errors.

Frequentrebalancesin a consumer group occur when:

Consumers are too slowand miss heartbeats (A), or

Instances crash or restart, triggering group membership changes (D)

FromKafka Consumer Group Coordination Docs:

“If a consumer fails to send a heartbeat in time (due to slowness or crash), it is removed from the group, causing a rebalance.”

Option B (partition mismatch) affectsload balancing, not rebalance frequency.

Option C (broker storage) doesn't trigger consumer rebalances.

To reset offsets for anexisting consumer group, you must use thekafka-consumer-groups.shtool with the --reset-offsets and --to-earliest flags.

FromKafka Consumer Group Tool Documentation:

“You can use the kafka-consumer-groups tool to reset offsets for a consumer group. This is required if the consumer has already committed offsets.”

Setting auto.offset.reset=earliest only worksif no committed offset exists.

Starting a new consumer with the same group won’t reset offsets.

Retention settings don’t affect committed offsets.

Log compactionensures that Kafka retains at least thelatest value per keyin a topic. Compaction happensin the backgroundand removes older records with the same keyin inactive log segments, not immediately.

From theKafka Documentation > Log Compaction:

“Kafka guarantees that thelast message for each keywill be retained in the log after compaction, even if earlier messages with the same key are deleted.”

So, D is correct. A is incorrect because compaction does not overwrite; it's a background process. B is incorrect—deletion is not immediate. C incorrectly suggests client-side hashmap behavior.

A:Kafka batches storecompression metadata in their headers, indicating how the batch was compressed.

B:Compression is controlled via theproducer config compression.type, which defaults to none. Producers explicitly set Gzip or other formats.

FromKafka Producer Docs:

“Compression is done by the producer, and the type is stored in the message batch metadata. The broker does not recompress.”

C is false: messages are serializedbeforecompression.

D is false: messages without keys are distributed acrossmultiple partitions.

For real-time applications, thekey metric to monitor is consumer lag— the difference between thelatest offset on the brokerand thelast committed offset by the consumer.

FromKafka Monitoring Guide:

“Consumer lag is the most important metric for real-time applications. It tells you how far behind the consumer is from the latest data in Kafka.”

B relates to replication, not client responsiveness.

C tracks group membership stability, not data delay.

D tracks input throughput, not processing latency.

A409 Conflictresponse in Kafka Connect REST API usually means:

The connector name provided already exists (i.e., a connector with that name is already deployed).

The Kafka Connect cluster is temporarily unavailable to accept new tasks due toa rebalance in progress.

From theConfluent Developer documentationand Kafka Connect REST API spec:

“409 – Conflict. The connector name is already in use or the worker group is in the process of rebalancing and cannot complete the request at this time.”

Option A is incorrect because invalid configurations return400 Bad Request, not 409.

Option B is invalid because Kafka Connect does not have a strict concept of "capacity" that prevents connector creation with a 409.

These three metrics are critical for cluster health:

OfflinePartitionsCount: Indicates partitions without active leaders — a sign of broker failure.

ActiveControllerCount: There should beexactly one active controller. A count ≠ 1 signals controller failure.

UncleanLeaderElectionsPerSec: Tracks leader elections where out-of-sync replicas were selected — risky for data loss.

FromKafka Monitoring Documentation:

“Offline partitions and unclean leader elections should trigger alerts. Also, ensure a single active controller is running.”

A is about topics pending deletion — not critical.

E is a per-topic record metric, not broker-level.