Calculate required Fabric Capacity

Hello dsj8wksnnckk,

Let me try help your. Is just a advise, sure?

To estimate the appropriate paid capacity when transitioning to Microsoft Fabric, it is essential to analyse the Fabric Capacity Metrics, focusing on CPU usage, memory consumption, and workload execution times (such as Dataflows Gen2 and Power BI). This analysis will help determine whether the current trial capacity is nearing its limits and which SKU would be most suitable.

Key Metrics to Monitor:

1. CPU Utilisation (%)
   • Review peak usage levels.
   • Frequent usage above 70–80% may indicate the need for a higher capacity tier.
2. Memory Usage (GB)
   • Assess both average and peak memory consumption.
   • If usage approaches the trial limit (F2 = 2 v-cores), consider scaling up.
3. Active Workloads
   • Identify which workloads are running (e.g., Power BI, Dataflows Gen2, Spark).
   • Each workload type consumes resources differently.
4. Queue Time
   • Monitor any delays in workload execution, which may suggest capacity saturation.
5. Concurrency
   • Evaluate the number of simultaneous operations. High concurrency may require additional v-cores.

Estimating the Ideal Capacity:

1. Export the capacity metrics data (via portal or API).
2. Analyse peak usage patterns across hours and days.
3. Compare against SKU limits:
   • For example: F2 (Trial) = 2 v-cores, F4 = 4 v-cores, F8 = 8 v-cores, etc.
   • Refer to official documentation: https://learn.microsoft.com/en-us/fabric/capacity/capacity-units
4. Simulate future scenarios:
   • Will the number of Dataflows increase?
   • Are more concurrent Power BI users expected?
   • Will additional workloads such as Lakehouse or Notebooks be introduced?

You can see more information here: Evaluate and optimize your Microsoft Fabric capacity - Microsoft Fabric | Microsoft Learn

If this response was helpful in any way, I’d gladly accept a 👍much like the joy of seeing a DAX measure work first time without needing another FILTER.

Please mark it as the correct solution. It helps other community members find their way faster (and saves them from another endless loop 🌀.