Capacity Utilization Reporting by Entity

Hi ashishdhakate03​,

The Microsoft Fabric Capacity Metrics app is very useful for monitoring capacity consumption, but if the requirement is to retain daily/weekly/monthly historical data by workspace, item, or other entity, you may need to persist the metrics yourself.

A practical approach is described here:

https://data-marc.com/2025/07/30/your-metrics-your-rules-extracting-and-storing-fabric-admin-and-capacity-metrics-data-in-fabric/

The approach uses the Capacity Metrics semantic model together with a composite model and Fabric/Semantic Link to extract the metrics and store them in a Lakehouse. Once the data is persisted, you can build your own historical Power BI report over it.

A bit couple of caveats

This approach relies on the Microsoft-provided Capacity Metrics semantic model, so I'd treat it as a practical/community solution rather than a guaranteed long-term API contract. The underlying Microsoft model can change, which could require adjustments to the extraction process.

I'd also recommend storing the extracted data incrementally in your own Lakehouse rather than continually overwriting it, so you can build a proper historical dataset.

So, if your requirement is "I need more than the native Metrics app's historical window and want to analyse capacity consumption by entity over weeks/months", this is a very reasonable approach to investigate.

For an enterprise implementation, I'd also validate the security implications, capacity overhead of the extraction process, and the supportability of depending on the Capacity Metrics semantic model before putting it into production.

From this, you can report:

• Daily CU consumption by workspace/item
• Weekly and monthly trends
• Top capacity consumers
• Interactive vs background consumption
• CU consumption by item type
• % of total CU consumption attributable to each entity
• Capacity utilization trends over time

Let me know if this helps.