Confluent Certified Developer for Apache Kafka v1.0
Question 1
Which configuration allows more time for the consumer poll to process records?
- A. session.timeout.ms
- B. heartbeat.interval.ms
- C. max.poll.interval.ms
- D. fetch.max.wait.ms
Answer : C
Question 2
Your application is consuming from a topic with one consumer group. The number of running consumers is equal to the number of partitions.
Application logs show that some consumers are leaving the consumer group during peak time, triggering a rebalance. You also notice that your application is processing many duplicates.
You need to stop consumers from leaving the consumer group.
What should you do?
- A. Reduce max.poll.records property.
- B. Increase session.timeout.ms property.
- C. Add more consumers instances.
- D. Split consumers in different consumer groups.
Answer : B
Question 3
A producer is configured with the default partitioner. It is sending records to a topic that is configured with five partitions. The record does not contain any key.
What is the result of this?
- A. Records will be dispatched among the available partitions.
- B. Records will be sent to partition '0'.
- C. An error will be raised and no record will be sent.
- D. Records will be sent to the least used partition.
Answer : A
Question 4
Which two producer exceptions are examples of the class RetriableException? (Choose two.)
- A. LeaderNotAvailableException
- B. RecordTooLargeException
- C. AuthorizationException
- D. NotEnoughReplicasException
Answer : AD
Question 5
Which partition assignment minimizes partition movements between two assignments?
- A. RoundRobinAssignor
- B. StickyAssignor
- C. RangeAssignor
- D. PartitionAssignor
Answer : B
Question 6
Which is true about topic compaction?
- A. When a client produces a new event with an existing key, the old value is overwritten with the new value in the compacted log segment.
- B. When a client produces a new event with an existing key, the broker immediately deletes the offset of the existing event.
- C. Topic compaction does not remove old events; instead, when clients consume events from a compacted topic, they store events in a hashmap that maintains the latest value.
- D. Compaction will keep exactly one message per key after compaction of inactive log segments.
Answer : D
Question 7
You create a producer that writes messages about bank account transactions from tens of thousands of different customers into a topic.
- Your consumers must process these messages with low latency and minimize consumer lag
- Processing these messages takes around six times longer than producing them
- Transactions for each bank account must be processed in order
You need to develop a strategy that minimizes consumer latency while maintaining scalability if the number of transactions increases.
Which strategy should you use?
- A. Use the timestamp of the message's arrival as its key.
- B. Use the bank account number found in the message as the message key.
- C. Use a combination of the bank account number and the transaction timestamp as the message key.
- D. Use a unique identifier such as a universally unique identifier (UUID) as the message key.
Answer : B
Question 8
You are sending messages to a Kafka cluster in JSON format and want to add more information related to each message. The information needed is:
- Format of the message payload
- Message creation time
- A globally unique identifier that allows the message to be traced as it flows through the system
Where should this additional information be set?
- A. Header
- B. Key
- C. Value
- D. Broker
Answer : A
Question 9
You have a Kafka consumer in production actively reading from a critical topic from which multiple other applications are consuming.
You have a new requirement to update the offset of your consumer to start reading from the beginning of the topic
Which action would you take?
- A. Temporarily configure the topic's 'retention.ms' parameter to '0' to empty the topic.
- B. Start a new consumer application with the same consumer group id.
- C. Update the consumer configuration by setting the 'auto offset reset property to 'earliest'.
- D. Update the consumer group's offset to the earliest position using the 'kafka-consumer-groups' CLI tool.
Answer : D
Question 10
Kafka producers can batch messages going to the same partition.
Which statement is correct about producer batching?
- A. Producers can only batch messages of the same size.
- B. Two or more broker failures will automatically disable batching on the producer.
- C. Producers have a separate background thread for each batch.
- D. Producers can include multiple batches in a single request to a broker.
Answer : D
Question 11
What is the default maximum size of a message the Apache Kafka broker can accept?
- A. 1 MB
- B. 2 MB
- C. 5 MB
- D. 10 MB
Answer : A
Question 12
You have a topic t1 with six partitions. You use Kafka Connect to send data from topic t1 in your Kafka cluster to Amazon S3. Kafka Connect is configured for two tasks.
How many partitions will each task process?
- A. 2
- B. 3
- C. 6
- D. 12
Answer : B
Question 13
This schema excerpt is an example of which schema format?
- A. Avro
- B. Protobuf
- C. JSON Schema
- D. YAML
Answer : B
Question 14
You have a consumer group with default configuration settings reading messages from your Kafka cluster. You need to optimize throughput so the consumer group processes more messages in the same amount of time.
Which change should you make to increase the consumption throughput?
- A. Remove some consumers from the consumer group.
- B. Increase the number of bytes the consumers read with each fetch request.
- C. Disable auto commit and having the consumers manually commit offsets.
- D. Decrease the session timeout of each consumer.
Answer : B
Question 15
You are writing a producer application and need to ensure proper delivery. You configure the producer with acks level to all.
Which two actions should you take to ensure proper error handling? (Choose two.)
- A. Use a callback argument in 'producer.send()' where you check delivery status.
- B. Check that 'producer.send()' retumed a 'RecordMetadata' object and is not null.
- C. Surround the call of 'producer.send()' with a try/catch block to catch 'KafkaException'.
- D. Check the value of 'ProducerRecord status()'
Answer : AC