To ensure that microservices can FAIL INDEPENDENTLY, we need to follow these principles:
π Each microservice should only facilitate a single business domain or function.
π Microservices should not share code or data.
π If you have to, rather violate the DRY principle than compromise on independence.
π Microservices should not communicate directly with each other over HTTP. Instead, they should make use of a message/event bus (message queue or broker, e. g. Apache Kafka)
π Each microservice should, as far as possible, be unaware of the existence of other microservices.
Top comments (0)