What is Apache Kafka and what problem does it solve?

Apache Kafka is a distributed event streaming platform: servers (brokers) and clients that capture events in real time, store them durably in topics for as long as you choose (the retention setting), and let applications process and route them. It solves the problem of moving continuous streams of data between applications. Producers and consumers are fully decoupled, topics are multi-subscriber, and events are not deleted after consumption, so they can be read as often as needed, which makes replay possible. Topics are split into partitions spread over several brokers and replicated, so if a server fails the others take over its work. Kafka's performance is effectively constant with respect to data size.

Expected key concepts: distributed, event streaming, durably, topics, decoupled, multi-subscriber, replay, partitions, replicated, retention