microservices-orchestration

Microservices Orchestration

This example showcases how Zeebe could orchestrate a payment microservice from within an order fulfillment microservice when Kafka is used as transport.

The Logger tasks are there only demo purposes to follow along on the console.

It leverages the connector's source to push a message onto Kafka whenever a payment is required, expecting some payment service to process it and emit an event (or send a response message) later on. This response is correlated to the waiting order fulfillment process using the connectors sink.

The example needs the payment service to be simulated, means you need to publish a record to the payment-confirm topic. You could do that using the kafka-console-producer:

{
  "eventType": "OrderPaid", 
  "orderId": 1,
  "amount": 4000
}

You can visualize the records published by the source connector using the kafka-console-consumer or simply Control Center. The records are published on the topic payment-request.

Prerequisites

• Install and run Kafka, Kafka Connect and Zeebe as described here. The following description assumes that you leverage a managed Zeebe Cluster in Camunda Cloud.

Running the example

Follow the following steps

Deploy process

You can use zbctl or the Camunda Modeler to deploy the process to Camunda Cloud.

Deploy connectors

Hint: If curl is not available, you can also use Control Center to create the connectors.

Make sure to configure them according to the following properties: source connector properties, sink connector properties

Now create the source connector:

curl -X POST -H "Content-Type: application/json" --data @payment-source.json http://localhost:8083/connectors

Next, create the sink connector:

curl -X POST -H "Content-Type: application/json" --data @payment-sink.json http://localhost:8083/connectors

Create a workflow instance

Now use the command line to to start a process instance, as you then can easily pass variables as JSON. Make sure to replace the Camunda Cloud connection information with your own:

zbctl --address 8fdfbf36-5c3a-49ff-b5c6-7057d396c88c.bru-2.zeebe.camunda.io:443 --clientId 6NlBrCXH5knkZsJod2xNaR~Z2Af45mYN --clientSecret TKJVqOUkauL-m93LjGaSlry6q.8~BsVIAiCFXsriK096qTEUbgGKw5q.SjE_YGhi create instance --variables "{\"orderId\": 1}" order

Replace the value of the orderId variable to change the correlation key.

Logger worker

Using Linux (or Mac) you can easily open a separate console, navigate to the root project directory, and run a worker directing all jobs to the console:

zbctl --address 8fdfbf36-5c3a-49ff-b5c6-7057d396c88c.bru-2.zeebe.camunda.io:443 --clientId 6NlBrCXH5knkZsJod2xNaR~Z2Af45mYN --clientSecret TKJVqOUkauL-m93LjGaSlry6q.8~BsVIAiCFXsriK096qTEUbgGKw5q.SjE_YGhi create worker --handler cat --maxJobsActive 1 payment-requested  & zbctl --address 8fdfbf36-5c3a-49ff-b5c6-7057d396c88c.bru-2.zeebe.camunda.io:443 --clientId 6NlBrCXH5knkZsJod2xNaR~Z2Af45mYN --clientSecret TKJVqOUkauL-m93LjGaSlry6q.8~BsVIAiCFXsriK096qTEUbgGKw5q.SjE_YGhi create worker --handler cat --maxJobsActive 1 payment-confirmed

If you do not want to start this worker, you can also simply wait 30 seconds for the time in BPMN to kick in and just skip the logging step.

Confirming order (Kafka Producer)

In order to simulate the external payment confirmation service, let's start a Kafka producer.

docker-compose -f docker/docker-compose.yml exec kafka  kafka-console-producer --request-required-acks 1 --broker-list kafka:19092 --topic payment-confirm

To confirm the order, we can write a record of the following format:

{"eventType": "OrderPaid", "orderId": 1, "amount": 4000}

Make sure to update the orderId to match the expected correlation key.