Study Confluent CCDAK Topics and Partitions: key concepts, common traps, and exam decision cues.
On this page
This lesson matters because many Kafka developer questions start with the storage and work-sharing model. The exam wants you to understand what a topic is, what a partition changes, what replication protects, and how a consumer group actually reads.
Model chooser
If the question is mostly about…
Strongest first focus
where records live
topic and partition model
who reads which work
consumer-group assignment
how consumer position is tracked
offsets
how copied data improves resilience
replication awareness
What the exam is really testing
If the scenario shows…
Strong reading
multiple consumers on one topic
group assignment and partition count matter
replay from a prior point
offsets are under test
copies of data on brokers
resilience awareness matters, even for developers
idle consumers
partition count may be the limiting factor
Common traps
Trap
Better rule
assuming every consumer in a group sees every record
a group shares partitions rather than duplicating work
confusing offset with partition
offset is a position inside a partition
treating replication as a read-scaling feature
replication is mainly a resilience and durability concern
Decision order that usually wins
Separate storage from work sharing: topics and partitions store records, consumer groups divide the reading work.
If the stem asks where a consumer is in the stream, think offset inside a partition, not a topic-wide position.
If consumers are idle, compare group size to partition count before blaming throughput.
If the question mentions replicas, stay in the resilience lane rather than drifting into read-scaling assumptions.