L3 - Default User Story (Required on creation):
In some cases of job failure, automated retries do not help (e.g. if there is an external cause of the failure that needs to be resolved first). It should be possible disable retries in the job executor in that case and immediately raise an incident.
Functional Requirements (Required before implementation):
- It is possible to mark specific job failures that should immediately set a job's retries to 0 (and thus raise an incident)
Technical Requirements (Required before implementation):
- The exact technical approach is to be specified
Limitations of Scope (Optional):
Hints (Optional):
This is the controller panel for Smart Panels app
[CAM-13295] I can throw an exception such that the job goes immediately to 0 retries
Hi fml2,
Thanks for your feedback. The approach you suggest makes sense, however the goal of this ticket is that the BPMN does not need to be modified in order to immediately raise an incident on an exception during job execution.
Cheers,
Thorben
OK, I can see the use case. If you implement it, please allow to specify the new retry schedule as the parameter in the exception. This new schedule would then be applied to this particular job instance. Then one could specify "no more retries" as a special case which would immediately result in an incident. But one could also say e.g. "one more time after 5 mins delay."
This ticket was migrated to github: https://github.com/camunda/camunda-bpm-platform/issues/2468. Please use this link for any future references and continue any discussion there.
If the application code can tell whether a retry would help, it could throw a BpmnError so that the process takes a special path. In that path, you'd create an incident.