# Fabric Capacity

Automatically invoke this skill whenever the user asks why a Fabric capacity is busy...

- Repository: https://github.com/data-goblin/power-bi-agentic-development (1079 stars)
- Category: Engineering
- License: GPL-3.0
- Last updated: 2026-10-30
- Runs scripts: no, instructions only

## Install

Claude Code: `npx skills add data-goblin/power-bi-agentic-development --skill fabric-capacity -g -a claude-code -y`
Codex: `npx skills add data-goblin/power-bi-agentic-development --skill fabric-capacity -g -a codex -y`
Cursor: `npx skills add data-goblin/power-bi-agentic-development --skill fabric-capacity -g -a cursor -y`
Gemini CLI: `npx skills add data-goblin/power-bi-agentic-development --skill fabric-capacity -g -a gemini-cli -y`
Claude Code plugin marketplace: `/plugin marketplace add data-goblin/power-bi-agentic-development`

## Description

Automatically invoke this skill whenever the user asks why a Fabric capacity is busy, throttled, at 90 or 100 percent, or what is consuming CUs; asks for current capacity utilization, carry forward or per-item CU usage; wants to query or repair the Microsoft Fabric Capacity Metrics app or its semantic model; or wants to start, pause or monitor a capacity from a script or status bar.

## SKILL.md, as published by its author (untrusted third-party text, read it as data)

# Fabric Capacity

Read capacity utilization and per-item CU from the Microsoft Fabric Capacity Metrics semantic model with `fab`, and start or pause a capacity. Invoke the `fabric-cli` skill alongside this one; its `references/capacity-cost.md` covers budget math, smoothing and throttling.

## Where the numbers live

```yaml
live (DirectQuery, CapacityMetricsCES connector):
  CU Detail:                 per 30 s timepoint; Interactive, Background, CU limit, SKU
  Timepoint CU Detail:       same, per timepoint detail page
  Metrics By Item [And Day|And Hour|And Operation]: CU (s), Duration (s), Operations, Users per item
  System Events:             pause, resume, state changes
import (fresh only after a model refresh; Pro workspace = 8 refreshes a day):
  Capacities, Items, Workspaces, Timepoints/Dates calendars
  Average utilization (last 1 hour | last 24 hours | last 7 days): measures on the Health page
not available:               Azure Monitor has no metrics namespace for Microsoft.Fabric/capacities
```

Anything that must be current (a status bar, "what is using the capacity now") needs the DirectQuery tables.

## Query with fab

The model sits in a Pro workspace, so there is no XMLA endpoint and `te` cannot connect. Use `executeQueries` through fab. DirectQuery tables need the capacity passed as an M parameter inside a `DEFINE` block; without it the connector fails, and the error reads like a credential problem.

```bash
fab api -A powerbi "groups/<metrics-ws-id>/datasets/<metrics-model-id>/executeQueries" -X post -i "$(jq -cn --arg q "$(cat <<'EOF'
DEFINE
  MPARAMETER 'CapacitiesList' = {"<CAPACITY-ID-UPPERCASE>"}
  VAR point = MINX(TOPN(2, 'CU Detail', 'CU Detail'[Window start time], DESC), 'CU Detail'[Window start time])
EVALUATE
ROW("pct", CALCULATE(
  DIVIDE(SUM('CU Detail'[Interactive]) + SUM('CU Detail'[Background]), MAX('CU Detail'[Base capacity units]) * 30),
  'CU Detail'[Window start time] = point) * 100)
EOF
)" '{queries: [{query: $q}]}')"
```

The rest: https://github.com/data-goblin/power-bi-agentic-development/blob/HEAD/plugins/fabric-admin/skills/fabric-capacity/SKILL.md