Lay out a day of Fabric workloads — pipelines, semantic model refreshes, Spark jobs, Power BI load — and see exactly where they collide. Find your peak, your overload windows and how much headroom is really left on an F64, F128 or F256 capacity.
Choose the F-SKU you run today, or the one you are sizing towards.
Add each workload with its start time, duration and CU demand — or start from a preset.
See the overload windows, what caused them, and which change clears them.
| Hour | Average CU | Peak CU | Peak utilization |
|---|
When each run starts and stops across the modeled day.
Share of the peak minute's demand.
Share of the day's total consumption, in CU-hours.
The demand curve does not change with the SKU — only the ceiling it is measured against.
| SKU | Capacity | Peak utilization | Overloaded for | Risk |
|---|
Demand is summed minute by minute and compared against the capacity's CU ceiling. No smoothing, bursting, carry-forward or workload-specific throttling is modeled, and concurrency is estimated from the timings you enter rather than measured CU telemetry. Treat the output as a modeled estimate for planning conversations — not as a substitute for the Fabric Capacity Metrics app.
We connect to your tenant, import real activity from the Fabric Capacity Metrics app, and tell you exactly which jobs to move — and whether you can drop a SKU.
BOOK A FREE CALL