Suppose that I have two pipelines - one which stores the incoming document in a specified queue and the other one which picks up the document from that queue and processes it. According to my understanding, the easiest way to turn these into ultra tasks would be:
- eliminate the first pipeline totally (as the queue is already present in ultra)
- eliminate the ‘picking up the document from the queue’ parts from the second pipeline
- call the second pipeline using the ‘pipeline execute’ snap from a new pipeline (if snaps are not compatible with ultra).
- construct ultra task using the new pipeline.
- use the second pipeline as ultra pipeline if all snaps are compatible.
Please let me know if my understanding is correct.
Also if some comparatively time consuming snaps are present in the ultra pipeline processing (assume tasks like 10+ SQLServer Execute, half a dozen REST Posts (ServiceNow create/update, other POSTs etc.), my understanding is that the caller will time out waiting for the response, but the ultra task will run to completion. In this case how can the caller know if it succeeded or not? In the original scenario that I’ve described, if the document has been put into the specified queue, it is assumed that it would be picked up and caller will get a response immediately.