Confluent CCAAK Brokers and Replication Guide

Study Confluent CCAAK Brokers and Replication: key concepts, common traps, and exam decision cues.

This lesson matters because many Kafka admin questions are really cluster-state questions in disguise. The exam wants you to know which broker role or partition state is actually responsible for what you are seeing.

Cluster-role chooser

If the stem is mostly about… Strongest first focus
metadata leadership and cluster state changes controller behavior
write target for one partition leader replica
replica catch-up health follower state and ISR membership
scaling read or write work partitions and replication design

What the exam is really testing

If the code or scenario shows… Strong reading
controller instability cluster coordination is under test
lagging followers replication health is under test
more consumers than parallel work partition count may be the constraint
leadership movement partition availability and safety are under test

Common traps

Trap Better rule
assuming all brokers are doing the same job at the same moment control and data roles still need to be distinguished
treating partitions like durability copies partitions drive parallelism; replicas drive resilience
blaming consumers for follower lag replication health is a broker-side concern first

Decision order that usually wins

  1. Separate cluster coordination from partition data flow: controller questions are not leader-replica questions.
  2. Ask whether the stem is about concurrency or resilience: partitions usually answer scale, replicas answer fault tolerance.
  3. If lag appears on followers, treat it as replication health first, not as a consumer tuning issue.
  4. If leadership is moving around, think availability and cluster stability before thinking application behavior.

Quiz

Loading quiz…
Revised on Sunday, May 10, 2026