Tools For Developing Zeebe Workflows and Workers using the Stream Consumer.

The following are packaged:

• Zeebe Broker
• Zeebe Simple Monitor (localhost:8082)
• Aether Stream Consumer (localhost:9013)
• Zeebe Tester (this app)

Zeebe Tester keeps a file system in sync with a Stream Consumer. While Tester is running, if you modify or add an artifact to the filesystem (./assets) it will be reflected automatically in the consumer. Likewise, if you save changes to a BPMN it will be updated in the broker.

Download the Zeebe Modeler if you want to edit the BPMNs: https://github.com/zeebe-io/zeebe-modeler

To start: scripts/up.sh

To wipe the broker and consumer (your artifact changes will stay unless you revert those with git) scripts/wipe.sh

Demo:

There are some interesting arifacts included in the package.

Start everything with:

• scripts/up.sh Visit the simple monitor.
• Zeebe Simple Monitor (localhost:8082)
• Look at the workflows (http://localhost:8082/views/workflows)
• We'll be drivin the http-getter instance.
• Click the "workflow-key" link to view it.

This workflow starts on a message sent to it giving it instructions, does work on an interval until it gets a second message telling it to stop.

The thing doing the work is a job/pipeline in the Stream Consumer. /assets/job/rest_handler.json & /assets/resource/pipeline/rest_handler.json which uses: a restcall and a zeebecomplete transformation.

Don't worry, you can modify any of these things, break it and put it back together.

This workflow is waiting for a message to kick it off. We have some samples in /messages

In worker-start.json

{
  "listener_name": "trigger-worker",
  "variables": {
    "worker_id":"one",
    "url": "http://www3.septa.org/hackathon/TransitViewAll/",
    "method": "get",
    "period": "R/PT10S"
  }
}

using httpie you can start it up with:

http --json POST http://localhost:9013/zeebe/send_message?id=default -a admin:password < ./messages/worker-start.json

Now look at instances again: http://localhost:8082/views/instances Click on the instance to watch it work. Take a look at the context and audit log. Sweet.

When you want to send the stop message (to the cancel-worker node on the diagram) we'll send the broker worker-cancel.json

{
  "listener_name": "cancel-worker",
  "correlationKey": "one"
}

again with httpie

http --json POST http://localhost:9013/zeebe/send_message?id=default -a admin:password < ./messages/worker-cancel.json