Hello, I am trying to understand the restrictions around the configuration of Message Correlation Keys. Could not find any explicit mention in the docs.
When I was testing something else I noticed that if I used the same value for Message Name, whenever I updated the Subscription Correlation Key of one message event with a value, all other messages had their Subscription Correlation Keys updated to the same value. This is a simplified version of a model that I used to reproduce it (using Camunda Desktop Modeler v. 5.48.0)
Ignore the model - it does nothing; it was just a way to test the updates in the Subscription Correlation Key. All Message Events have the Message Name = âMessage_Testâ and the value I set to Subscription Correlation Key in any of the Message Events is replicated to the others.
I thought âOkay, there must be a restriction and the same process instance cannot have subscriptions with same Message Name and different Subscription Correlation Keys.â But then I remembered another test I did with a multi-instance Receive Task.
In this test I pass a collection of different strings to the Receive Task that also had Message Name = âMessage_Testâ. Each string is used as Subscription Correlation Key for each instance of the Receive Task, resulting in 3 simultaneous subscriptions:
Message Name = "âMessage_Testâ, Subscription Correlation Key = â0123â
Message Name = "âMessage_Testâ, Subscription Correlation Key = â4567â
Message Name = "âMessage_Testâ, Subscription Correlation Key = â8901â
I ran this process and it worked perfectly. So, the engine can have simultaneous subscriptions of messages with the same Message Name and different Subscription Correlation Keys.
My doubt is: if the engine supports different Subscription Correlation Keys for the same Message Name why the Modeler wonât allow me to configure different values?
PS. Tried the Web Modeler and it seems to behave the same way.

