Streaming & Event-Driven Lessons

Exactly-once semantics in Kafka explained: idempotent producers, transactions, and committing consumer offsets inside the producer transaction to make read-process-write loops exactly-once.

Streaming and event-driven architecture from first principles: logs, topics, partitions, delivery guarantees, event time vs processing time, and how to design sound async pipelines.

Watermarks and event-time windowing explained: how a watermark estimates completeness, advances by the min rule, triggers windows, handles late data, and why an idle input stalls a pipeline.