Why Chargeback Matters

In a shared-capacity model like Microsoft Fabric, multiple teams and workloads consume from the same CU pool. Without a chargeback mechanism, there is no accountability for consumption — leading to over-provisioning, wasted spend, and friction between business units and central IT.

A well-designed chargeback strategy provides:

• Cost transparency: Every team sees what they consume and what it costs
• Accountability: Business units are incentivized to optimize their workloads
• Budget alignment: IT can distribute infrastructure costs fairly across the organization
• Capacity governance: Data-driven decisions about scaling, reservations, and capacity topology

Chargeback vs. Showback

The Chargeback Process

Implementing Fabric chargeback is a multi-step process that combines organizational decisions, Azure billing data, and the Chargeback App:

There is no one-size-fits-all approach to chargeback. What works for one organization may not be right for another. Your strategy must answer two fundamental questions:

Key Strategic Decisions

1. Handling Unused Compute

Within the capacity model there will almost always be unused compute — even with the benefits of smoothing (just like unused minutes on a phone contract). Capacities are priced such that 100% utilization isn't needed to be cost-effective. However, how you attribute unused compute is critical:

2. Handling Reservations

Azure Reservations provide approximately 40% discounts on pay-as-you-go prices but require a 1-year or 3-year commitment. Key considerations:

• Reservations apply opportunistically to any capacity within scope — you cannot pin a reservation to a specific capacity
• If only certain business units are willing to commit long-term, you may need to separate reserved and PAYG costs and charge back separately
• With multiple capacities of different sizes sharing a reservation, finding total cost per capacity requires calculating blended rates (reserved + PAYG)
• Use Azure Cost Management chargeback for reservations to identify amortized costs per capacity

3. Administration Uplift

If capacities are created and managed by a central team, you may want to charge an uplift on Fabric compute usage to cover the costs of that central IT team (FTE costs, tooling, on-call support). This is a percentage added on top of the raw Azure cost.

4. Scale-Up, Pause/Restart & Overage Costs

Dynamic capacity management (scaling up for peak loads, pausing during off-hours, or using capacity overage) affects the total bill. Decide upfront whether these costs are absorbed centrally or distributed to the teams that triggered the demand.

Strategy Decision Matrix

What the App Does

The Microsoft Fabric Chargeback App is a Power BI app available from AppSource that connects to your Fabric capacity data and provides breakdowns of CU consumption. It helps you answer: "What percentage of my capacity did each team/user/workspace consume?"

What the App Does NOT Do

• It does not tell you how much money to charge each workspace, business unit, or user
• It does not calculate total costs — you must get those from Azure Cost Management
• It does not handle reservation cost allocation — use Azure Cost Management amortized costs
• It does not apply organizational cost models (uplift, rate cards, etc.)
• It shows proportional usage against used capacity, not available capacity — important for unused compute calculations

Installation & Setup

Prerequisites

• You must be an admin of at least one capacity in your organization
• Install the app in a workspace with a Pro license to avoid throttling due to capacity overutilization

Installation Steps

1. Go to AppSource → Microsoft Fabric Chargeback and select Get it now
   Or in Power BI service: Apps → Get apps → search "Microsoft Fabric" → select Chargeback → Get it now
2. Sign in with your Microsoft account and select Install
3. Open the installed app and select Connect your data
4. Configure the connection:

   Field	Value	Notes
   UTC Offset	-12 to 14 (use .5 for half-hour zones)	Your org's standard UTC offset
   Days Ago to Start	14 or 30 days	Historical data to load
   Advanced → Auto Refresh	On (default) or Off	Midnight auto-refresh

5. Set Authentication method to OAuth2 and Privacy level to Organizational
6. Select Sign in and continue — initial data load may take a few minutes

App Visuals & Navigation

• Chargeback page: Main view with tabs for Workspace, Item, Domain, and Users. Shows top N entries by utilization percentage.
• Drill-through: Right-click any workspace, item, or domain to drill into detail pages showing utilization breakdowns, user details, and daily trends.
• Export Data page: A matrix visual for exporting utilization and user details across all capacities. Use slicers to filter and select columns for export.
• Utilization (CU) by date: Daily utilization trend chart. Useful for identifying peak usage patterns.

Considerations & Limitations

• Data is not real-time — refreshed daily (manual refresh available from the workspace)
• Operations not associated with a user (or by service principals) show as "Power BI Service"
• When "Show user data" is disabled in admin settings, users appear as "Masked user" (counted as a single user)
• Export may fail with "visual has exceeded available resources" — apply filters to reduce data volume
• Not supported in government clouds or environments using private links
• The semantic model is for use only within the provided reports — modifications are unsupported

A complete chargeback workflow combines two data sources:

Calculating Chargeback Amounts

Team Charge = Total Capacity Cost × (Team CU Usage / Total CU Used)

Example:
  Total monthly cost (from Azure): $8,400
  Finance team CU usage:           40% (from Chargeback App)
  Planning team CU usage:          35%
  Operations team CU usage:        25%

  Finance charge:    $8,400 × 0.40 = $3,360
  Planning charge:   $8,400 × 0.35 = $2,940
  Operations charge: $8,400 × 0.25 = $2,100

Used Cost = Total Capacity Cost − Compute Pool Cost (from Azure meter)

Team Charge = Used Cost × (Team CU Usage / Total CU Used)

Example:
  Total monthly cost:        $16,800
  Compute Pool (unused) cost: $3,400
  Used cost:                 $13,400

  Business Unit A (60% of used): $13,400 × 0.60 = $8,040
  Business Unit B (40% of used): $13,400 × 0.40 = $5,360
  Central IT absorbs:            $3,400 (unused)

Team Charge = (Total Cost × Team Usage %) × (1 + Admin Uplift %)

Example with 10% uplift:
  Base charge for Team A: $3,360
  With 10% uplift:        $3,360 × 1.10 = $3,696

Organizational Readiness

Technical Best Practices

• Enforce domain assignments: If using domain-based chargeback, ensure every workspace is assigned to a domain. Unassigned workspaces appear as "No domain" and can't be attributed. Build a validation process to catch gaps.
• Maintain workspace mappings: If using workspace-based attribution, keep an up-to-date mapping of workspaces to teams/owners. Cross-check monthly against the Chargeback App export.
• Automate the export-to-bill pipeline: Manually combining Azure Cost Management data with Chargeback App exports is error-prone. Consider building a Power BI report or Power Automate flow that joins the two data sources automatically.
• Monitor for anomalies: Set up alerts for unexpected spikes in CU consumption using the Capacity Metrics App. Investigate before bills go out to avoid disputes.
• Separate dev/test from production: Use different capacities (or at minimum, different workspaces with clear domain assignments) for dev/test and production. This simplifies chargeback and avoids mixing experimental costs with production.
• Understand billing for dynamic operations: Scale-ups, pause/restarts, and capacity overage all affect the total bill. Document how each is handled in your chargeback policy.

Maturity Model