The Kafka Real-Time Stream Processing works with the real-time data at the rate of message generation. The message processing is continuous and without blocking. Many business processes are also continuous and don't need a response to start the processing. Streaming Data is essential for such business processes as reporting suspicious transactions or tracking mail deliveries.
Any messages appearing in the application can be published in a special Kafka topic. It might be any document in the ERP system or any user action on the website, including clicks, adding to favorites, adding/removing from the Cart, filling in forms, viewing pages (and a visit depth), can be sent and distributed by the specified Kafka topics. So, other topics (consumers) can subscribe to the necessary topics for various purposes, such as monitoring, analysis, reporting, personalization, etc.
All Kafka messages are added to the log (saved to a drive) and stored until the moment of clearing the log history, which a user sets in advance. Due to this feature, Kafka can be used as a reliable source of data (unlike RabbitMQ, which deletes messages right after delivery).