There should be an example that shows how a "Job Retry Profile" can be applied to certain job declarations.
There could be different profiles for different external systems:
External System 1ERP System External System 2CRM System
A profile defines a default retry time cycle, which should be applied when "talking" with that external system. In case of "External System 1" it could be "R5/PT10M" and in case of "External System 2" it could be "R10/PT5M" etc.
To avoid to set the retry time cycle and to add a asynchronous continuation (asyncBefore or asyncAfter) flag for each task where the process "talks" with one of the external task, the user can use a property to set a "profile" inside the bpmn model. The property can than be used inside a process engine plugin to set default retry time cycle behavior depending on the given profile.