Apache Kafka is a popular choice for powering data pipelines. Apache kafka added kafka stream to support popular etl use cases. KSQL makes it simple to transform data within the pipeline, readying messages to cleanly land in another system. KSQL is the streaming SQL engine for Apache Kafka. It provides an easy-to-use yet powerful interactive SQL interface for stream processing on Kafka, without the need to write code in a programming language such as Java or Python. KSQL is scalable, elastic, fault-tolerant, and real-time. It supports a wide range of streaming operations, including data filtering, transformations, aggregations, joins, windowing, and sessionization.

https://docs.confluent.io/current/ksql/docs/index.html

对于etl系统来说，Rabbitmq并不是一个受欢迎的选择，它更适合那些需要简单的消息传递系统和更低吞吐量的系统。

Scaling both is hard in a distributed fault tolerant way but I'd make a case that it's much harder at massive scale with RabbitMQ. It's not trivial to understand Shovel, Federation, Mirrored Msg Queues, ACK, Mem issues, Fault tollerance etc. Not to say you won't also have specific issues with Zookeeper etc on Kafka but there are less moving parts to manage. That said, you get a Polyglot exchange with RMQ which you don't with Kafka. If you want streaming, use Kafka. If you want simple IoT or similar high volume packet delivery, use Kafka. It's about smart consumers. If you want msg flexibility and higher reliability with higher costs and possibly some complexity, use RMQ.

在以下情况使用RabbitMQ:

你不需要处理大数据，你更喜欢一个方便的内置UI来监控 不需要自动复制队列 消息没有多个订阅者——因为不像Kafka是一个日志，RabbitMQ是一个队列，消息一旦被消费和确认到达就会被删除 如果您有要求使用通配符和正则表达式的消息 如果定义消息优先级很重要

简而言之: RabbitMQ适用于简单的用例，数据流量低，具有优先级队列和灵活的路由选项。 对于海量数据和高吞吐量使用Kafka。

我能想到的唯一好处是事务性功能，其余的都可以用Kafka来完成

你们忘记的一个关键区别是RabbitMQ是基于推的消息系统，而Kafka是基于拉的消息系统。这在消息传递系统必须满足具有不同处理能力的不同类型的消费者的场景中非常重要。使用基于Pull的系统，消费者可以根据自己的能力消费，而推送系统将推送消息，而不管消费者的状态如何，从而将消费者置于高风险之中。

我知道有点晚了，也许你已经间接地说过了，但是，Kafka根本不是一个队列，它是一个日志(就像上面有人说的，基于民意调查)。

简单来说，当你更喜欢RabbitMQ(或任何队列技术)而不是Kafka时，最明显的用例是:

You have multiple consumers consuming from a queue and whenever there is a new message in the queue and an available consumer, you want this message to be processed. If you look closely at how Kafka works, you'll notice it does not know how to do that, because of partition scaling, you'll have a consumer dedicated to a partition and you'll get into starvation issue. Issue that is easily avoided by using simple queue techno. You can think of using a thread that will dispatch the different messages from same partition, but again, Kafka does not have any selective acknowledgment mechanisms.

你能做的最多的就是像那些家伙一样，试着把Kafka转换成一个队列: https://github.com/softwaremill/kmq

雅尼克