Copyright (c) 2022 by Fred George
@author Fred George [email protected]
Licensed under the MIT License; see LICENSE file in root.
A Kotlin implementation of Rapids/Rivers framework using RabbitMQ for the event bus
This framework models the Rapids/Rivers/Pond metaphor first proposed by Fred George. It strives to present a relatively simple choreography scheme to allow tiny MicroServices to interact with minimal coupling.
Originally formulated for a workshop conference in Bergen, Norway, and it has been re-implemented several times, including at NAV (Norwegian Welfare Association) in 2019 where it is now being used extensively in the organization, and at Orn Software in 2022 (in C#).
This project is built with:
- IntelliJ IDEA 2023.3.6 (Ultimate Edition)
- Java 17.0.7 Oracle
- Kotlin 1.9.23
- JUnit 5.10.2
- Gradle 8.7
- RabbitMQ amqp driver 5.18.0
- Jackson JSON 2.17.0
So open up the project, and let the dependencies resolve themselves. You can execute the tests to ensure everything is working correctly if you are uncertain.
Now you are ready to run the sample MicroServices.
An event bus acts as a rapids, an undifferentiated stream of all messages. Access to the rapids is through an adapter; for RabbitMQ, the adapter is the RabbitMqRapidsConnection.
Messages on the event bus are captured into packets. In Kotlin and Java, you can inspect the received fields through the APIs of Packet. You can also create new packets, and add properties to them that will be rendered into JSON automatically on publishing to the rapids.
Using a set of rules, a subset of the rapids is created as a river. Services are attached to a river, and will receive only packets that conform to the rules.
Two Kotlin sample services are provided:
- Need in Kotlin: service generates a stream of messages on the event bus
- Examine this code for how to create and send a Packet
- Monitor in Kotlin: service logs to the console all messages on the bus
- Examine this code for how to attach to the rapids via a river
- Also note the comments for how to set up rules for the packets you want to process
Two corresponding Java sample services (JMonitor and JNeed) are also provided with the same purposes as the Kotlin versions.
Bring up an instance of RabbitMQ, preferably with the management addition to allow browser inspection of what is happening.
Run the Monitor service. Be sure to add the startup parameters identifying the RabbitMQ IP address (localhost if running locally) and port number (default is 5672). You should see a console log announcing its start.
Next, run the Need service. Again, IP and port are required parameters. Need should generate a message very five seconds. Monitor should show this message.
If this is all working, you are ready to write your own, new services!