Microsoft Fabric Capacity Governance: F SKUs, Bursting and Debt
Diesen Beitrag auf Deutsch lesen
Fabric capacity units form a shared compute pool. F SKUs, burst limits, smoothing and the Capacity Metrics app shape performance and cost governance.
TL;DR
Microsoft Fabric capacity units form a shared compute pool across workspaces. F SKUs can be paused and resumed, while bursting creates carry forward debt. Interactive work is smoothed over a few minutes, background workloads over 24 hours. The Capacity Metrics app helps admins track utilization, debt and top consumers, with alerts at roughly 80 percent.
Original on M365 FM Podcast. Read the original · suggested by Mirko Peters
This is our own summary, not a republication or full translation.
Governance takeaway
- Admins/CoE: Monitor shared capacity with the Microsoft Fabric Capacity Metrics app, including utilization, carry forward debt and top-consuming models, pipelines or notebooks.
- Admins/CoE: Review F SKU sizing and inherited P SKUs, because shared workloads and legacy capacity arrangements affect performance and migration planning.
- Leadership/Business: Treat capacity as a cross-department resource, because intensive work in one workspace can affect unrelated reports and cloud cost control.
Frequently asked questions
How does Microsoft Fabric shared capacity affect different workspaces?
All workspaces draw from the same capacity pool, so an intensive notebook job in one department can affect reports in another.
What does the Microsoft Fabric Capacity Metrics app monitor?
The app shows utilization trends, carry forward debt and top-consuming semantic models, pipelines or notebooks, helping capacity admins investigate performance.
How does carry forward debt affect Fabric report performance?
Bursting beyond the purchased capacity creates debt that future compute cycles must repay, potentially throttling interactive reports after the original workload runs.