Kafka FAQs: Zookeeper Dependency, Retention Policies, Cleanup Rules, Performance Bottlenecks, and Cluster Best Practices

152. Can Kafka run without Zookeeper? Why?

Kafka cannot run without Zookeeper because it uses Zookeeper to manage and coordinate its broker nodes.

153. How many data retention strategies does Kafka have?

Kafka has two data retention strategies: retention by expiration time and retention by total message size.

154. If both a 7‑day and a 10 GB retention limit are set and the data reaches 10 GB on the fifth day, how does Kafka handle it?

Kafka will trigger data cleanup as soon as either condition is met, so it will delete data once the size limit is reached, regardless of the time limit.

155. What situations can cause Kafka to slow down?

CPU performance bottlenecks

Disk I/O bottlenecks

Network bottlenecks

156. What should be considered when using a Kafka cluster?

Having too many brokers can increase replication latency and reduce overall throughput; it is recommended to keep the cluster size ≤ 7.

Prefer an odd number of brokers so that a majority can survive failures, improving fault tolerance.

(End)