A message processor is used to deliver messages that have been temporarily stored in a message store. This approach is useful for serving traffic to back-end services that can only accept messages at a given rate, whereas incoming traffic to the ESB arrives at different rates. You use the Store mediator to store messages in the message store, and then you use a message processor to deliver messages from the message store to the back-end service at a given rate. Using message processors and message stores allows you to implement different messaging and integration patterns.
Multiple message processors can use the same message store. For example, in a clustered environment, each of the nodes might have an instance of the same message processor, each of which would connect to the same message store and evenly consume messages from it. The message store acts as a manager of these consumers and their connections and ensures that messages are processed by only one message processor, preventing message duplication. You can further control which nodes a message processor runs on by specifying pinned servers, so that you might have each message processor run on one node only, as shown in the following diagram.
The ESB ships with the following message processor implementations:
You can also implement your own message processor by implementing the MessageProcessor interface and adding the message processor to the configuration.
Message Processor Configuration
Sampling and custom processors
Message Processor does not work in tenant mode. It is a limitation in the current implementation.