Skip to main content

Cluster for Apache Kafkas

Kafka cluster operations are the work of running brokers reliably. A cluster has several brokers with partition leaders and followers, replication, and a controller that, in current versions, uses KRaft instead of ZooKeeper. Operators manage retention, log compaction, partition reassignment, monitoring and upgrades.

Stage: Operations · Part of the Apache Kafka roadmap · Last reviewed 19 November 2026

What to learn in Cluster

Cluster interview questions

What is consumer lag and why monitor it?

Consumer lag is the difference between the latest offset in a partition and the group's committed offset. Growing lag means consumers are not keeping up, so data is processed late. It is a primary health signal and often triggers scaling or investigating slow processing.

What is an in-sync replica?

An in-sync replica is a follower that has caught up with the partition leader within a configured time. Only in-sync replicas can be elected leader without risking data loss. If followers fall behind, the in-sync set shrinks, which affects acks=all writes.

Where Cluster fits in the Apache Kafka roadmap

Cluster belongs to the Operations stage (advanced level). Run it.

Back to the full Apache Kafka roadmap