P — Performance & Capacity¶
How platforms remain healthy as they grow
Power Platform's licensing and capacity model is nuanced, and operational management of cost and capacity requires deliberate attention. The good news: the Admin Center and CoE tooling provide most of what you need if configured correctly.
Power Platform Implementation¶
Capacity and Cost Management¶
- Power Platform Admin Center — Capacity page — Displays Dataverse database, file, and log storage consumption per environment. Set up storage alerts before environments approach limits; storage overages incur additional cost and can block solution imports.
- Licensing inventory — Maintain a current inventory of Power Apps per-app, per-user, and Microsoft 365-included licences. The CoE Starter Kit's licence management components surface which users have which licences and where premium licences are in use.
- Pay-as-you-go (PAYG) metering — For environments with unpredictable usage, consider linking environments to an Azure subscription for PAYG licensing. This provides cost transparency and prevents licence constraint from blocking legitimate usage.
- AI Builder credits — Track AI Builder credit consumption via the Admin Center. AI Builder is credit-based and can be depleted unexpectedly if flows using AI models are widely shared.
- Dataverse for Teams vs. full Dataverse — Be deliberate about which environments use full Dataverse vs. Dataverse for Teams. The latter has capacity limits and feature restrictions that can create migration costs later.
Performance and Forecasting¶
- Use Admin Center analytics and Application Insights data to establish performance baselines for critical applications (load time, flow execution duration, API call volumes)
- Model capacity growth: as adoption increases, storage and API call volumes grow. Build this into your annual capacity planning cycle
- For high-volume flows (thousands of runs per day), monitor Power Automate throughput against your licence tier's daily limits. Cloud flows have per-user and per-flow run limits that can cause throttling at scale
Key Signals to Monitor¶
| Signal | Tool | Action Threshold |
|---|---|---|
| Dataverse storage | Admin Center — Capacity | Alert at 80% of allocated storage |
| API call volumes | Admin Center — Analytics | Review when approaching daily limits |
| Flow run failures | Flow run history / App Insights | Alert on failure rate > defined threshold |
| AI Builder credits | Admin Center | Alert at 70% of monthly allocation |
| Inactive premium licences | CoE Starter Kit | Monthly review and reallocation |
Getting Started¶
- Baseline storage consumption per environment in the Admin Center
- Deploy the CoE Starter Kit licence management components for visibility of licence usage
- Enable Application Insights for production apps to capture performance metrics
- Set up capacity alerts in the Admin Center before environments approach limits
- Include capacity forecasting in your annual planning cycle