I have a dataflow that hits mostly an Azure SQL database but there are a couple other sources as well.
All of my transformation steps occur in the dataflow.
This dataflow is on a Premium Per User workspace and refreshes every half hour and typically takes about 7 minutes to do so.
I have a dataset that is solely derived from this dataflow. No transformation occurs in the dataset. It simply just loads the tables from the dataflow.
The dataset is in the same premium per user capacity scheduled to refresh every 30 minutes. On a normal refresh it takes roughly 12 minutes to complete. However, I'm frequently running into scenarios (at least once a day sometimes more) where the refresh hangs for 3-4 hours. It typically doesn't fail but just hangs.
I don't understand why the dataflow where the transformation is ocurring is rock solid and never hangs, but the dataset gives me so many issues.
Any ideas?
