Starting on 6/17/25, we have been experiencing intermittent refresh failures on random reports. The same error message occurs each time:
"The refresh operation failed because it took too long to complete. Consider reducing the size of your dataset or breaking it up into smaller datasets."
- We have a Premium P1 SKU
- Primary source is an Azure SQL server
- Affected models vary from .05 to 2 GB in size
- Affected models normally refresh within 1-8 minutes
- Affected models have been normally refreshing for up to a few years until 6/17/25 without experiencing this error
- No recent changes have been made to Power Query for affected models
- This failure occurs in between 2-4 minutes of a refresh start
- This failure only occurs between 4-6a PT for scheduled refreshes
- This failure never seems to occur on the same report twice
- All reports affected refresh multiple times per day, and never experience this error twice in a row
- After failing, a manual refresh immediately after always completes successfully; the error cannot be manually reproduced
- I've already explored common causes of this actual error and none apply to these semantic models or our configuration
I'm thinking this is a bug introduced on PBI Service's end and the error message being sent is a false positive. Possibly a new error related to Azure SQL concurrent connections for the mashup engine? Has anyone else seen this start occurring recently?
