Menu
Back to Discussions

when should you add a message queue?

ive been working on a few microservices lately, and the communication between them is getting a bit messy. we're currently using direct HTTP calls for most inter-service communication, but it feels like we're heading towards a situation where a failure in one service can cascade and bring down others. ive seen message queues like RabbitMQ and Kafka used in many architectures, and they seem to offer solutions for decoupling and async processing. at what point in a projects lifecycle or complexity should one seriously consider introducing a message queue? what are the common tipping points beyond just asynchronous needs?
36 comments

Comments

Loading comments...