Reflex starting fabric items – latency when triggering Fabric items

[KimTutein]

I have been playing around with Reflex setting up a reflex triggering a fabric pipline on a blob created event on a storage account using private endpoint. I am running on a F64 capacity.

The triggering works but doing the last day of testing the latency has been extremely different. Sometimes it has started the pipeline within seconds but sometimes it takes +5 minutes. Looking into the Reflex select/detect I can see that the event is triggered within a few seconds – however the “act part” is triggering much later.

Anyone had experience similar and can advise – I would expect it to start the pipeline within seconds?

[KimTutein]

Hi @lbendlin 

Thank you for your reply - that makes sense.

I hope Microsoft will improve this as I this as a way into triggering pipelines in Fabric from both users (that do not want to waite 5 minutes to see if there file is read into the system) and triggering Fabric load from external systems. If I compare with Microsoft Synapse pipelines where we have triggers from blob events they work great - se need the same in Fabric for us to move there.

[lbendlin · Accepted solution]

Reflexes currently have no QoS whatsoever, they are treated as low priority compared to all other activities on the capacity.  The Data Activator product team is aware and are hopefully considering changes.