What does Kafka guarantee when using a consumer group with three partitions and three consumers?

When using a consumer group with three partitions and three consumers, Kafka guarantees:

• Each partition will be assigned to exactly one consumer in the group. Thus, all three consumers will read data in parallel, but without overlaps across partitions.
• Message order within a single partition is preserved. The consumer reading a specific partition receives messages in the same order they were written.
• Load balancing is ensured. If there are fewer consumers than partitions, some consumers will read multiple partitions; if there are more, the extra consumers will be idle.

Therefore, with an equal number of partitions and consumers, each consumer processes a unique partition, ensuring parallel processing without duplication and guaranteeing order within partitions.