Advanced
Apache Kafka
A log, and everything that follows from it.
Topics, partitions and the ordering they give, producers and acknowledgements, consumer groups, rebalancing, offsets and commits, retries and dead-letter topics, schema evolution, exactly-once semantics, and Kafka Streams.
8 lessons written2 modules~2h reading
After this course you can
- Choose a partition key and explain the ordering and throughput consequences
- Configure a producer for durability and a consumer for at-least-once without duplicates causing harm
- Handle poison messages with retries and a dead-letter topic
- Explain what exactly-once means in Kafka and what it does not cover
Curriculum
8 lessons · outlined lessons show their plan01
Topics, partitions and orderingThe log abstraction, partitions as the unit of parallelism, keys and ordering guarantees, and replication and ISR.15 minProducers and acknowledgementsacks, idempotent producers, batching and linger, and the message that was "sent" and never arrived.14 minConsumer groups and rebalancingGroup coordination, partition assignment, the rebalance storm, cooperative rebalancing, and why one consumer per partition is the ceiling.14 minOffsets and delivery guaranteesCommit strategies, at-most-once and at-least-once, the idempotent consumer, and the message processed twice after a crash.14 minThe log
02
Retries and dead-letter topicsBlocking versus non-blocking retries, retry topics with backoff, poison pills, and what to do with a dead letter.13 minSchema evolutionAvro, Protobuf and JSON Schema with a registry, compatibility modes, and the consumer that broke on a renamed field.13 minKafka with SpringKafkaTemplate, @KafkaListener, concurrency, error handlers, and the tests that use an embedded broker.12 minKafka StreamsStream and table, stateful operations, windows, joins, and exactly-once processing for the pipelines that transform events.16 min