Microsoft Fabric Deployment Options: A Successful Fabric Deployment Is Not a Working Solution
·1188 words·6 mins
The deployment succeeded, everything is green, and someone opens the report to find yesterday’s data. The tool did what it was asked to do. We asked too little of the release. Part two of my Fabric deployment series goes through what each item brings with it. A lakehouse moves as metadata and shortcut definitions, without its tables or files. The active value set in a variable library is workspace state, so the pipeline comparison won’t even show that test and production use different values. Role assignments, refresh schedules and data source credentials for semantic models don’t come along. And promoting workspace content doesn’t update a classic Power BI app. A green status isn’t my finish line. Someone from the client team should open the result and check that the data is current and points where it should.