Tell me about the message brokers you have worked with.

In my practice, I have worked with several message brokers, mainly RabbitMQ and Apache Kafka.

RabbitMQ was used for tasks where reliable message delivery and support for various protocols (AMQP) are important. For example, in microservice architecture for asynchronous task processing and load balancing. An example in Go using the github.com/streadway/amqp library:

conn, err := amqp.Dial("amqp://guest:guest@localhost:5672/")
if err != nil {
    log.Fatal(err)
}
channel, err := conn.Channel()
// ...
channel.Publish("exchange", "routing.key", false, false, amqp.Publishing{
    ContentType: "text/plain",
    Body:        []byte("Hello RabbitMQ"),
})

Apache Kafka was used for processing large data streams with high throughput and scalability. I used the confluent-kafka-go library for producing and consuming messages. Kafka is well-suited for event sourcing and building systems with real-time event processing.

I also have experience with Redis Streams for lightweight queues and Pub/Sub when simplicity and low latency are required.

The choice of broker depends on requirements for reliability, speed, scalability, and infrastructure complexity.