View Full Confluent CCDAK Exam Dumps and Practice Test Dumps

 

Question 21

What does ISR mean in Kafka replication?

1. In-Sync Replicas
2. Internal Storage Records
3. Indexed Stream Routes
4. Instance State Registry

Correct Answer: 1

Explanation:

ISR stands for In-Sync Replicas. These are replicas that are considered sufficiently caught up with the partition leader according to Kafka’s replication rules. The ISR set is important for durability because producer acknowledgments and settings such as min.insync.replicas can depend on the availability of in-sync replicas. A replica that falls too far behind may be removed from the ISR until it catches up. CCDAK architects should understand ISR behavior when designing highly available Kafka clusters because broker failures, network delays, and overloaded replicas can affect both durability and write availability.

Question 22

Which Kafka architecture replaces ZooKeeper for cluster metadata management?

1. Legacy replication mode
2. KRaft mode
3. MirrorMaker mode
4. Connect distributed mode

Correct Answer: 2

Explanation:

KRaft is Kafka’s architecture for managing cluster metadata without requiring Apache ZooKeeper. It uses Kafka’s own metadata quorum and controller mechanisms to manage cluster state. This simplifies the overall architecture by removing the separate ZooKeeper dependency used by older Kafka deployments. KRaft controllers participate in maintaining metadata information and coordinating cluster operations. The change is architecturally significant because deployment, administration, scaling, and failure handling differ from ZooKeeper-based environments. CCDAK candidates should understand KRaft because modern Confluent Kafka deployments increasingly use this architecture for cluster metadata management.

Question 23

Which feature prevents duplicate producer writes after retries?

1. Idempotent producer
2. Consumer buffering
3. Topic compaction
4. Replica reassignment

Correct Answer: 1

Explanation:

Kafka’s idempotent producer capability helps prevent duplicate records caused by producer retries. The producer and broker cooperate using producer identifiers and sequence information so retried writes can be recognized appropriately. This is particularly valuable when network failures create uncertainty about whether a previous request succeeded. Idempotence does not mean that an application can ignore all duplicate-processing concerns in downstream systems, but it provides an important Kafka-level guarantee for producer retries. CCDAK architects should understand how idempotence interacts with acknowledgments, retries, batching, and transactional processing when designing reliable event publication workflows.

Question 24

Which setting limits the minimum ISR count required for writes?

1. replica.fetch.max.bytes
2. unclean.leader.election.enable
3. min.insync.replicas
4. segment.bytes

Correct Answer: 3

Explanation:

min.insync.replicas specifies the minimum number of replicas that must remain in sync for a producer using an appropriate acknowledgment requirement to successfully write data. For example, a topic with a replication factor of three might require at least two in-sync replicas. If the ISR falls below that threshold, writes requiring the configured durability level can fail rather than accepting data with insufficient replication. This setting helps balance availability against durability. CCDAK architects should configure it together with replication factor and producer acknowledgment settings rather than treating it as an isolated parameter.

Question 25

What Kafka mechanism records the current leader’s generation?

1. Consumer epoch
2. Leader epoch
3. Producer generation
4. Topic revision

Correct Answer: 2

Explanation:

A leader epoch identifies a generation of leadership for a Kafka partition. Whenever leadership changes, the epoch advances. Kafka uses this information to help brokers and clients recognize stale leadership information and correctly handle changes in partition leadership. This is important during broker failures, elections, recovery, and replica synchronization. A leader epoch is different from a record offset because an offset identifies a position within a partition log, while the epoch identifies a particular leadership generation. Understanding leader epochs helps CCDAK architects reason about failover behavior and why clients can safely recover from changing partition leaders.

Question 26

Why is rack awareness used when assigning Kafka replicas?

1. Separate replicas physically
2. Increase schema versions
3. Reduce consumer groups
4. Disable partition leaders

Correct Answer: 1

Explanation:

Rack awareness helps Kafka distribute replicas across different failure domains, such as racks, availability zones, or comparable infrastructure boundaries. If replicas are concentrated in the same physical location, a failure affecting that location could remove multiple copies simultaneously. Distributing replicas improves resilience against localized infrastructure failures. Rack awareness does not alter schemas or consumer group membership. CCDAK architects should incorporate infrastructure topology into replica placement decisions, especially for production environments spanning multiple availability zones. The goal is to reduce correlated failures while maintaining practical network and latency characteristics.

Question 27

Which setting controls whether an out-of-sync replica may become leader?

1. auto.leader.rebalance.enable
2. unclean.leader.election.enable
3. replica.lag.time.max.ms
4. leader.imbalance.check.interval.seconds

Correct Answer: 2

Explanation:

unclean.leader.election.enable controls whether Kafka may elect a replica that is not currently in the in-sync replica set when no ISR is available. Allowing an unclean election can improve availability but may result in data loss because the selected replica can be missing records that existed on the previous leader. Disabling unclean elections prioritizes data consistency over availability when all in-sync replicas are unavailable. This is an important architectural trade-off. CCDAK candidates should evaluate business requirements for durability and availability before choosing the appropriate behavior.

Question 28

Which Kafka feature groups records into producer batches?

1. Record timestamp
2. Partition reassignment
3. Producer batching
4. Consumer assignment

Correct Answer: 3

Explanation:

Producer batching allows multiple records destined for the same partition to be accumulated and sent together. Batching can improve throughput by reducing the number of individual requests and making network and broker processing more efficient. Producer settings such as batch.size and linger.ms influence batching behavior. Larger batches can improve efficiency but may increase latency or memory usage depending on workload characteristics. CCDAK architects should evaluate message rate, record size, latency requirements, and broker capacity when tuning producer batching rather than optimizing only for maximum throughput.

Question 29

Which producer setting specifies the maximum time to wait before sending a batch?

1. linger.ms
2. compression.type
3. retries
4. client.id

Correct Answer: 1

Explanation:

linger.ms controls how long a Kafka producer may wait for additional records to arrive before sending a partially filled batch. Allowing a small delay can produce larger batches and improve throughput, particularly when records arrive at a moderate rate. Setting the value too high can increase end-to-end latency. batch.size determines the target batch capacity rather than the waiting duration. CCDAK architects should tune batching parameters according to application latency requirements and traffic patterns. A high-throughput workload may benefit from modest batching delays, while latency-sensitive applications may prefer shorter waiting periods.

Question 30

Which producer option controls the compression algorithm used for records?

1. delivery.timeout.ms
2. compression.type
3. max.in.flight.requests.per.connection
4. client.dns.lookup

Correct Answer: 2

Explanation:

compression.type determines the compression algorithm used by the Kafka producer for record batches. Common choices include algorithms such as gzip, snappy, lz4, and zstd, depending on the Kafka version and environment. Compression can reduce network traffic and storage requirements, although it introduces CPU processing costs. The best choice depends on workload characteristics, infrastructure resources, and performance goals. CCDAK architects should evaluate compression together with batch sizes and message rates because compression efficiency generally improves when batches contain sufficient data for the algorithm to process effectively.

Question 31

Which producer setting limits the total memory available for buffering unsent records?

1. request.timeout.ms
2. metadata.max.age.ms
3. buffer.memory
4. retry.backoff.ms

Correct Answer: 3

Explanation:

buffer.memory controls the amount of memory available to the Kafka producer for buffering records that have not yet been transmitted to brokers. When the producer’s buffer becomes full, calls that need additional memory may block until space becomes available or until the relevant timeout is reached. This setting is therefore connected to producer throughput, batching, broker responsiveness, and application memory usage. It should not be confused with broker-side memory settings. CCDAK architects should size producer buffering according to expected traffic bursts while avoiding excessive memory consumption that could destabilize application processes.

Question 32

Which producer setting defines the overall delivery deadline for a record?

1. delivery.timeout.ms
2. metadata.max.idle.ms
3. receive.buffer.bytes
4. reconnect.backoff.max.ms

Correct Answer: 1

Explanation:

delivery.timeout.ms establishes the upper time boundary for completing a producer send operation, including retries and waiting for acknowledgment. It provides an overall delivery deadline rather than controlling only one network request. This setting works alongside parameters such as request timeout and retry behavior. If the producer cannot successfully deliver a record within the configured period, the send operation can fail. CCDAK architects should choose a value consistent with application latency expectations, broker responsiveness, retry requirements, and failure-recovery behavior so temporary network problems do not create inappropriate application failures.

Question 33

What does the Kafka controller primarily manage?

1. Application serialization
2. Consumer payloads
3. Cluster metadata operations
4. External database writes

Correct Answer: 3

Explanation:

The Kafka controller manages important cluster-level metadata operations. Depending on the Kafka architecture, controller responsibilities include coordinating partition leadership and maintaining cluster metadata state. In KRaft-based deployments, controller nodes participate in a metadata quorum rather than relying on ZooKeeper. Controllers do not serialize application payloads or directly manage external database writes. Understanding controller responsibilities is important for CCDAK architecture because controller availability and quorum behavior influence cluster management, partition leadership, and metadata operations. Architects should distinguish controller responsibilities from those handled by brokers, producers, consumers, and Connect workers.

Question 34

Which setting determines the maximum size of a producer request?

1. auto.offset.reset
2. max.request.size
3. fetch.max.wait.ms
4. enable.auto.commit

Correct Answer: 2

Explanation:

max.request.size limits the maximum size of a request sent by a Kafka producer. It can restrict the amount of data included in a single request and therefore influences how large batches can become. The setting should be considered alongside broker and topic message-size limits because the effective maximum payload is constrained by the configuration across the producer and broker path. auto.offset.reset controls consumer offset behavior, while fetch-related parameters affect consumers. CCDAK architects should ensure producer message-size settings are compatible with application requirements and broker-side limits.

Question 35

Which Kafka setting controls the number of requests allowed in flight per connection?

1. fetch.min.bytes
2. max.poll.interval.ms
3. max.in.flight.requests.per.connection
4. heartbeat.interval.ms

Correct Answer: 3

Explanation:

max.in.flight.requests.per.connection controls how many unacknowledged requests can exist simultaneously on a producer connection. Increasing the value can improve throughput by allowing more requests to be outstanding, while lower values can reduce concurrency. When retries are involved, the setting can also affect ordering considerations depending on producer idempotence and other configuration. CCDAK architects should evaluate this setting with throughput, latency, retry behavior, and ordering requirements. It is particularly important when designing producers that must balance efficient network utilization against predictable request sequencing.

Question 36

Which Kafka setting controls automatic consumer offset commits?

1. enable.auto.commit
2. fetch.max.bytes
3. isolation.level
4. partition.assignment.strategy

Correct Answer: 1

Explanation:

enable.auto.commit determines whether a Kafka consumer automatically commits offsets at intervals controlled by the consumer configuration. Automatic commits can simplify consumer applications, but they may not align with business processing guarantees if records require successful downstream processing before an offset should advance. Applications needing tighter control can disable automatic commits and explicitly manage offsets. Other settings serve different purposes: isolation.level affects transactional record visibility, while assignment strategy controls partition distribution. CCDAK architects should select offset management behavior based on the application’s processing semantics and tolerance for duplicate or missed processing.

Question 37

What does Kafka’s isolation.level=read_committed hide from consumers?

1. Aborted transactional records
2. Ordinary retained events
3. Compressed batches
4. Replicated partitions

Correct Answer: 1

Explanation:

With isolation.level=read_committed, consumers read only records from committed transactions and do not expose records belonging to aborted transactions. This behavior is important when applications use Kafka transactions and require consumers to process only successfully committed event sequences. By contrast, read_uncommitted allows consumers to see transactional records regardless of whether their transactions ultimately commit. CCDAK architects should understand this distinction when designing end-to-end exactly-once processing or transactional pipelines. Producer transactions, consumer isolation, and downstream processing behavior must be considered together to achieve the intended application semantics.

Question 38

Which Kafka capability atomically writes records to multiple partitions?

1. Consumer rebalance
2. Transactional producer
3. Schema compatibility
4. Partition reassignment

Correct Answer: 2

Explanation:

Kafka transactions allow a producer to atomically publish records to multiple partitions. This is useful when several output records must either all become visible or none should be considered committed. Transactions can also integrate offset commits with produced records for appropriate consume-transform-produce workflows. A transactional producer requires suitable configuration and application behavior to achieve the intended semantics. Consumer rebalancing, schema compatibility, and partition reassignment address different concerns. CCDAK architects should evaluate transaction overhead and operational complexity against the application’s consistency requirements before selecting transactional processing.

Question 39

Which broker setting controls the number of log segments retained before rolling by size?

1. segment.bytes
2. min.insync.replicas
3. replica.fetch.wait.max.ms
4. broker.rack

Correct Answer: 1

Explanation:

segment.bytes determines the approximate size at which Kafka rolls a log segment and begins writing to a new segment. Kafka stores partition logs as a sequence of segment files rather than as one continuously growing file. Segment boundaries affect operations such as retention, deletion, and log management because Kafka generally applies certain policies at the segment level. Smaller segments can allow more granular cleanup but may increase file-management overhead. CCDAK architects should consider workload volume, retention behavior, disk characteristics, and operational requirements when evaluating segment sizing.

Question 40

Which Kafka policy removes records based on both time and available disk space?

1. compact
2. delete
3. compact,delete
4. archive

Correct Answer: 3

Explanation:

The compact,delete cleanup policy combines log compaction with ordinary deletion-based retention. Compaction retains the latest record for each key, while deletion removes older log segments according to retention settings. This combination is useful when an application wants current keyed state while also allowing obsolete historical data to age out. A compact policy alone focuses on compaction, while delete uses standard retention behavior. CCDAK architects should choose cleanup policies based on whether the topic represents an event history, a current-state changelog, or a combination of both requirements.