Capacity Monitoring
Overview of the Capacity & Cost Monitoring module — view active capacities, optimization alerts, and navigate to detailed capacity insights.
Overview
The Capacity & Cost Monitoring module lets you monitor your Microsoft Fabric and Power BI Premium capacities directly from Soterre PBI Analyzer. When you open this module, you see a dashboard with your active capacities, optimization alerts, and quick-access cards for each capacity.
Prerequisites
Before using Capacity Monitoring:
- Connect to Power BI Service — you need at least one tenant configured with valid credentials. See Add a New Tenant.
- Install Fabric Capacity Metrics app — required for CU consumption data and real-time metrics. See Step 6: Install Fabric Capacity Metrics App.
- Provide Metrics Workspace ID and Dataset ID — enter these when adding your tenant. Without them, CU metrics and some alerts will be unavailable.
- Add the service principal to the Metrics workspace — at least Member role (Viewer is not enough for DAX queries).
Required Licenses
| Feature | Minimum License |
|---|---|
| Basic capacity info | Power BI Pro |
| CU metrics and alerts | Fabric Capacity (F SKU) or Power BI Premium (P SKU) |
The Capacities Dashboard
When you open the Capacity Monitoring page, you see:
Tenant Selector
A dropdown in the top-right corner lets you choose which tenant to view. If you only have one tenant configured, it is selected automatically.
Summary Cards
Two cards at the top of the page:
| Card | What It Shows |
|---|---|
| Active Capacities | Count of capacities in Active state, with a sub-label showing how many are inactive |
| Optimization Alerts | Count of the cost-estimate SKU right-sizing suggestions (the dashboard's only suggestion source). Turns orange if the count is above zero; the sub-label breaks it into critical and high-priority counts. In the current build these SKU suggestions are effectively never generated (see the note under Optimization Suggestions), so this count is normally 0 |
Capacity Cards
Below the summary, each capacity appears as a clickable card showing:
- Name and a colored state badge (green = Active, orange = Inactive/Suspended, red = Deleting/Deleted)
- SKU (e.g., F64, P1) and Region
- Memory Limit in GB with a badge showing the SKU tier (Fabric, Premium, Embedded, Azure, or Trial)
Click any card to open its detail page.
Optimization Suggestions
Suggestions are rendered as a list of cards. Each suggestion includes:
- Severity icon — Critical (red), High (orange), Medium (yellow), Low (blue)
- Title and description of the issue
- Recommendation — what to do about it
- Affected items — up to 3 affected names shown as chips, plus a "+N more" chip when there are more
The dashboard list only shows SKU right-sizing suggestions
On this dashboard, the Optimization Suggestions list and the Optimization Alerts count are both driven by the cost-estimate endpoint, which only ever emits an SKU downsize suggestion (severity Low). That suggestion is gated on a capacity's utilization being below 30%, but utilization is currently a hardcoded placeholder of 50%, so the SKU suggestion is effectively never generated. As a result, the dashboard's suggestions list and alert count are normally empty — even for a capacity at high memory usage.
When there are no suggestions, the Optimization Suggestions section is not rendered at all (there is no empty-state message on the dashboard).
Memory, refresh, and failure suggestions
The memory-, refresh-, and failure-based suggestions below are generated per capacity and appear on each capacity's detail page, not on this dashboard:
| Suggestion | Severity | Trigger |
|---|---|---|
| Excessive refresh frequency | Medium | Any dataset refreshes more than 8 times per day |
| Recent refresh failures | High | Any dataset had a failed refresh in the past 7 days |
| Memory approaching limit | Medium | Memory usage ≥ 80% |
| Memory critically high | Critical | Memory usage ≥ 95% |