**Video about Kafka Architecture

**Video about Kafka Architecture

**https://www.youtube.com/watch?v=-AZOi3kP9Js

One of the best lectures about Kafka architecture: straight to the point, great to listen to at 1.75x speed, enough to get a complete understanding of the technology

I will be writing an article on implementing EDA and RPC on Kafka, but in brief:

EDA – each consumer group will receive an instance of a message from a topic, so it's enough to distribute different applications across different consumer groups and you will have the ability to process Events (the only difficult moment is choosing the number of partitions, but I would take a number equal to half of the largest consumer group)

RPC – creates a response topic with 1 partition and a unique name (among all instances) from the sender, adds a unique message ID (correlationId) and the response topic name to the Kafka message headers, and sends it to the topic from which RPC requests are read (it's better to have a separate topic for calls, but you can also specify the call name as the key of a message in a large topic). The consumer reads the message, forms a response, and sends it to where you specified in the headers.

I will describe examples of implementation in the article (and there will also be a similar article on RabbitMQ)

P.S.

I forgot to mention that the simplest way to deploy Kafka on K8s is strimzi.io (I strongly recommend it)