Message correlation debugging - broker logging in general


I am trying to debug a problem with a workflow not moving forward when receiving an event message.
I am stepping through code with my ide and can see that the message is sent, but is there a way to adjust the broker logs to see what is happening there? (I am using docker-compose-up.)


In case someone is interested in the actual problem, it was a syntax problem in the modeler.
The “Message name” field should not have quotes around it (I’m quite sure that have seen such examples from Googling), like ‘MESSAGE_NAME’ (without quotes). It is apparently interpreted as a string with the the “=” sign for a variable. And of course the “Message correlation key” should be like ‘=orderId’ (without quotes).
I’m still interested in learning more about the logging though.

Check return code. It empty if everything ok, and has error message, if no process found to correlate message or two or many processes found

In the broker, we don’t have logging to track the workflow processing. Instead, you can use an exporter to consume the log stream. It contains different records to track the workflow processing. For messages, you should look for records of type: message, message-subscription, and workflow-instance-subscription.

You can also use the Zeebe simple monitor: It shows additional information that might help :slight_smile:

No, this is not true. You don’t get an error response if the message could not be correlated, or if it correlates to multiple instances.

Thank you, Philipp.

Yes, I use Simple Monitor to debug message correlation during development. This is one cool feature that is has for development that Operate doesn’t have yet.