Forum Discussion
Semantic Model Views and User Count
- 12 days ago
Hi Ashish,
The Usage Metrics report is built from report render events only, so Analyze in Excel traffic will never show there because AIX connects straight to the model XMLA endpoint and does not open a report page.
Two reliable ways to get what you want:
1. Fabric Workspace Monitoring (preview, requires an F capacity). Enable it on the workspace, select Semantic Model as the item type, and every DAX query that hits the model lands in an Eventhouse (KQL database). The SemanticModelLogs table gives you the user UPN, timestamp, application/source, duration, and the DAX text. This includes Analyze in Excel, SSMS, Tabular Editor, report queries, and any XMLA client. Point a Power BI report at that KQL DB and you have the user count, active users, and query volume you listed.
2. Power BI admin activity log (works at Pro too). Either from the Purview portal or scripted with Get-PowerBIActivityEvent, filter on the AnalyzedByExternalApplication operation. Each entry has UserId, DatasetId, and workload, so you can group by user per model. The retention is 30 days from the API and 90 days from the UI, so if you need longer history, dump it nightly into a Lakehouse or SQL table.
If the model is on PPU or a Premium/Fabric capacity, Workspace Monitoring is the much better experience because the data is granular and already structured for querying.
If this helped, a thumbs up and marking it as the accepted solution would be appreciated.
Best,
Shai Karmani
Hi ashishkr10006,
I second Shai_Karmani's suggestion of GetActivityEvents. One thing to note is the API only stores 28 days worth of data, so if you want to see history further back than that you need to create a process that will save the results from the API to a lakehouse or warehouse.
https://learn.microsoft.com/en-us/rest/api/power-bi/admin/get-activity-events
I suggest to all fabric admins to set up a process like this that runs daily to save the previous days audit data somewhere safe so that they can start to build a full history of what people are doing in their Fabric environment.