Application Integration
Amazon MQ
Managed message broker for Apache ActiveMQ Classic and RabbitMQ that handles their installation, operation, and maintenance. It lets you migrate applications that already use standard messaging protocols and APIs (JMS, AMQP, MQTT, STOMP) without rewriting their messaging code.
Exam phrases that point here
«migrate applications that already use a message broker to the cloud without rewriting the messaging code»
Amazon MQ is a managed broker for Apache ActiveMQ and RabbitMQ that speaks the same standard APIs and protocols; usually all it takes is pointing the applications at the new broker's endpoint.
When not: If the application is new or can be rewritten, AWS recommends considering Amazon SQS and Amazon SNS: they are lighter-weight, fully managed, and scale almost without limit.
«applications communicate using standard protocols such as JMS, AMQP, MQTT, or STOMP»
Amazon MQ supports APIs such as JMS and NMS and protocols such as AMQP, STOMP, MQTT, and WebSocket, so existing clients keep working.
When not: If the scenario only asks for a queue or notifications and does not mention a dependency on a standard protocol or API, the expected answer is usually Amazon SQS or Amazon SNS.
«stop managing the broker on Amazon EC2 (patching, upgrades, and failure recovery)»
Amazon MQ provisions the broker's infrastructure and software and handles upgrades, security patches, failure detection, and recovery.
When not: If custom plugins or deep customization of the broker are needed, the FAQ indicates that it is better to keep managing it on EC2.
«highly available message broker across Availability Zones»
For ActiveMQ there is the active/standby deployment (two brokers in two AZs with shared storage on Amazon EFS); for RabbitMQ, a three-node cluster.
When not: A single-instance broker is one broker in one AZ, without the redundancy of those deployments. If compatibility with a broker is not needed, SQS is a fully managed service that is usually simpler.