Kafka interview questions

1. A.a copy of a partition stored on another broker; one is the leader and the others follow for fault tolerance
2. B.a Kafka producer that assigns sequence numbers so retries do not create duplicate records on a partition
3. C.the position of a record within a partition
4. D.the unit of parallelism and ordering within a topic

Answer + AI explanation with Pro

1. A.Kafka consumer group
2. B.Kafka partition
3. C.Single Message Transform
4. D.Producer batching

Answer + AI explanation with Pro

1. A.Kafka partition — the default Kafka producer partitioner for null-key records that fills one partition per batch before switching, improving batching and throughput over round-robin
2. B.Kafka partition — the Kafka topic config setting how long messages are kept before deletion regardless of consumption, independent of any consumer offsets
3. C.Kafka partition — a set of consumers where each partition is read by exactly one member
4. D.Kafka partition — the unit of parallelism and ordering within a topic

Answer + AI explanation with Pro

1. A.the set of in-sync replicas that are sufficiently caught up to the leader and eligible to be elected leader
2. B.a class implementing Kafka's Partitioner interface to control which partition a record lands in beyond the default key-hash or sticky behavior
3. C.a Kafka producer that writes to multiple partitions atomically using a transactional ID so consumers see all or none
4. D.a set of consumers where each partition is read by exactly one member

Answer + AI explanation with Pro

1. A.subject compatibility mode
2. B.Kafka exactly-once
3. C.Kafka transactions
4. D.Kafka consumer group

Answer + AI explanation with Pro

1. A.Kafka consumer group — a Kafka producer that assigns sequence numbers so retries do not create duplicate records on a partition
2. B.Kafka consumer group — a set of consumers where each partition is read by exactly one member
3. C.Kafka consumer group — achieved via idempotent producer + transactions + read_committed
4. D.Kafka consumer group — the framework of reusable source and sink connectors that streams data between Kafka and external systems without custom producer or consumer code

Answer + AI explanation with Pro

1. A.reassigning partitions when a consumer joins or leaves the group
2. B.retention that keeps only the latest value per key
3. C.a Kafka producer that writes to multiple partitions atomically using a transactional ID so consumers see all or none
4. D.the default Kafka producer partitioner for null-key records that fills one partition per batch before switching, improving batching and throughput over round-robin

Answer + AI explanation with Pro

1. A.Producer batching
2. B.Broker
3. C.Rack awareness
4. D.consumer group rebalancing

Answer + AI explanation with Pro

1. A.consumer group rebalancing — the gap between the latest offset on a partition and the offset a consumer group has committed
2. B.consumer group rebalancing — a Kafka producer that assigns sequence numbers so retries do not create duplicate records on a partition
3. C.consumer group rebalancing — reassigning partitions when a consumer joins or leaves the group
4. D.consumer group rebalancing — retention that keeps only the latest value per key

Answer + AI explanation with Pro

1. A.grouping multiple records into one request, controlled by batch.size and linger.ms, to raise throughput at the cost of latency
2. B.a service storing versioned Avro, Protobuf, or JSON schemas and enforcing compatibility so producers and consumers evolve message formats safely
3. C.a Kafka server that stores topic partitions, serves producer writes, and answers consumer fetches
4. D.the producer setting requiring all in-sync replicas to acknowledge a write

Answer + AI explanation with Pro

1. A.retention.ms
2. B.Topic
3. C.acks=all
4. D.Consumer lag

Answer + AI explanation with Pro

1. A.acks=all — the producer setting requiring all in-sync replicas to acknowledge a write
2. B.acks=all — the framework of reusable source and sink connectors that streams data between Kafka and external systems without custom producer or consumer code
3. C.acks=all — the Schema Registry setting (BACKWARD, FORWARD, FULL) governing which schema changes are allowed for a subject before registration is rejected
4. D.acks=all — a service storing versioned Avro, Protobuf, or JSON schemas and enforcing compatibility so producers and consumers evolve message formats safely

Answer + AI explanation with Pro

1. A.achieved via idempotent producer + transactions + read_committed
2. B.a Kafka feature that spreads partition replicas across racks or availability zones to survive zone failures
3. C.the framework of reusable source and sink connectors that streams data between Kafka and external systems without custom producer or consumer code
4. D.a Kafka retention policy that keeps only the latest record per key, useful for changelog and state topics

Answer + AI explanation with Pro

1. A.Kafka offset
2. B.Kafka exactly-once
3. C.Kafka partition
4. D.Broker

Answer + AI explanation with Pro

1. A.Kafka exactly-once — the Kafka topic config capping the size of a partition log on disk, triggering deletion of the oldest segments once exceeded
2. B.Kafka exactly-once — a Kafka producer that writes to multiple partitions atomically using a transactional ID so consumers see all or none
3. C.Kafka exactly-once — achieved via idempotent producer + transactions + read_committed
4. D.Kafka exactly-once — the Schema Registry setting (BACKWARD, FORWARD, FULL) governing which schema changes are allowed for a subject before registration is rejected

Answer + AI explanation with Pro

1. A.retention that keeps only the latest value per key
2. B.a Kafka feature that spreads partition replicas across racks or availability zones to survive zone failures
3. C.the Kafka Connect-based replication tool that copies topics, configs, and consumer offsets across clusters with offset translation for disaster recovery and migration
4. D.the unit of parallelism and ordering within a topic

Answer + AI explanation with Pro

1. A.Rack awareness
2. B.log compaction
3. C.subject compatibility mode
4. D.acks=all

Answer + AI explanation with Pro

1. A.log compaction — a class implementing Kafka's Partitioner interface to control which partition a record lands in beyond the default key-hash or sticky behavior
2. B.log compaction — retention that keeps only the latest value per key
3. C.log compaction — a Kafka feature that spreads partition replicas across racks or availability zones to survive zone failures
4. D.log compaction — a Kafka broker limit (bytes per second) applied per client or user to throttle producers and protect the cluster from being overwhelmed

Answer + AI explanation with Pro

1. A.grouping multiple records into one request, controlled by batch.size and linger.ms, to raise throughput at the cost of latency
2. B.a Kafka record with a key and a null value that signals deletion of that key in a compacted topic, removed after delete.retention.ms
3. C.the position of a record within a partition
4. D.the Schema Registry setting (BACKWARD, FORWARD, FULL) governing which schema changes are allowed for a subject before registration is rejected

Answer + AI explanation with Pro

1. A.Kafka offset
2. B.sticky partitioner
3. C.custom partitioner
4. D.Topic

Answer + AI explanation with Pro

1. A.Kafka offset — the unit of parallelism and ordering within a topic
2. B.Kafka offset — a Kafka retention policy that keeps only the latest record per key, useful for changelog and state topics
3. C.Kafka offset — a lightweight per-record transformation applied inside Kafka Connect to mask, route, or reshape messages as they flow through a connector
4. D.Kafka offset — the position of a record within a partition

Answer + AI explanation with Pro

1. A.a Kafka guarantee combining idempotent producers and transactions so consume-process-produce pipelines neither lose nor duplicate records
2. B.a named, append-only log of records in Kafka that is split into partitions for parallelism and scale
3. C.the Schema Registry setting (BACKWARD, FORWARD, FULL) governing which schema changes are allowed for a subject before registration is rejected
4. D.the Kafka topic config capping the size of a partition log on disk, triggering deletion of the oldest segments once exceeded

Answer + AI explanation with Pro

1. A.Log compaction
2. B.Consumer lag
3. C.Topic
4. D.Kafka offset

Answer + AI explanation with Pro

1. A.Topic — a lightweight per-record transformation applied inside Kafka Connect to mask, route, or reshape messages as they flow through a connector
2. B.Topic — a Kafka producer that assigns sequence numbers so retries do not create duplicate records on a partition
3. C.Topic — a named, append-only log of records in Kafka that is split into partitions for parallelism and scale
4. D.Topic — retention that keeps only the latest value per key

Answer + AI explanation with Pro

1. A.the consumer isolation level that returns only records from committed transactions and skips aborted ones, required for exactly-once reads
2. B.a copy of a partition stored on another broker; one is the leader and the others follow for fault tolerance
3. C.the gap between the latest offset on a partition and the offset a consumer group has committed
4. D.the position of a record within a partition

Answer + AI explanation with Pro

1. A.Schema Registry
2. B.Single Message Transform
3. C.Replica
4. D.Transactional producer

Answer + AI explanation with Pro

1. A.Replica — the framework of reusable source and sink connectors that streams data between Kafka and external systems without custom producer or consumer code
2. B.Replica — the producer setting requiring all in-sync replicas to acknowledge a write
3. C.Replica — the unit of parallelism and ordering within a topic
4. D.Replica — a copy of a partition stored on another broker; one is the leader and the others follow for fault tolerance

Answer + AI explanation with Pro

1. A.achieved via idempotent producer + transactions + read_committed
2. B.the Kafka topic config setting how long messages are kept before deletion regardless of consumption, independent of any consumer offsets
3. C.the set of in-sync replicas that are sufficiently caught up to the leader and eligible to be elected leader
4. D.a copy of a partition stored on another broker; one is the leader and the others follow for fault tolerance

Answer + AI explanation with Pro

1. A.Consumer lag
2. B.Exactly-once semantics
3. C.ISR
4. D.producer quota

Answer + AI explanation with Pro

1. A.ISR — a Kafka feature that spreads partition replicas across racks or availability zones to survive zone failures
2. B.ISR — the Kafka topic config setting how long messages are kept before deletion regardless of consumption, independent of any consumer offsets
3. C.ISR — the set of in-sync replicas that are sufficiently caught up to the leader and eligible to be elected leader
4. D.ISR — the producer setting requiring all in-sync replicas to acknowledge a write

Answer + AI explanation with Pro