e2e-fabric-cost-estimation

Estimates Microsoft Fabric capacity cost by profiling workloads and fetching live Azure pricing.

1.1k|301|Updated Feb 17, 2026
One-click install
npx skills add https://github.com/microsoft/skills-for-fabric --skill e2e-fabric-cost-estimation
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: e2e-fabric-cost-estimation
Source: https://github.com/microsoft/skills-for-fabric/tree/main/plugins/fabric-skills/skills/e2e-fabric-cost-estimation
Command: npx skills add https://github.com/microsoft/skills-for-fabric --skill e2e-fabric-cost-estimation

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

Migrating Spark, SQL, Power BI, and Real-Time workloads to Microsoft Fabric requires accurate capacity sizing and cost forecasting, but pricing is region-specific, meter names are non-obvious, and hardcoded rates go stale. This Skill profiles source workloads, maps them to Fabric Capacity Units, and computes live SKU and billing-mode costs.

Core Features & Use Cases

  • Workload-to-CU Mapping: Converts Databricks clusters, Synapse DWU/Spark pools, Power BI P-SKUs, and multi-cloud sources (AWS, GCP, Snowflake) into Fabric CU demand with a concurrent-demand sizing algorithm.
  • Live Pricing Enforcement: Mandates fetching the Azure Retail Prices API for PAYG, Reserved Instance, Autoscale Spark, OneLake storage, and egress rates before any dollar figure is presented.
  • Billing Mode Strategy: Compares Reserved Instance vs. Pay-As-You-Go break-even points and Autoscale Billing for Spark vs. capacity-billed Spark.
  • Use Case: A team migrating from Databricks and Power BI Premium P2 asks which Fabric SKU to buy; the Skill profiles their cluster hours and P-SKU utilization, fetches live East US pricing, and produces a monthly cost worksheet comparing F128 RI against PAYG plus Autoscale Spark.

Quick Start

Ask the assistant to estimate the monthly Fabric capacity cost for migrating your Databricks and Power BI workloads, specifying your Azure region and workload profile.

Frequently Asked Questions about e2e-fabric-cost-estimation

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I estimate Microsoft Fabric capacity cost before a migration?

Profile your source workloads (cluster hours, DWU-hours, P-SKU utilization), map them to Fabric Capacity Units, then fetch live per-CU rates from the Azure Retail Prices API for your region. Multiply by SKU CU counts and 730 hours for PAYG monthly figures.

How do Databricks vCores map to Fabric Capacity Units?

The documented conversion is 1 Fabric CU equals 2 Spark vCores, so a node's CU equivalent is its vCores divided by 2. These cross-platform mappings are unvalidated heuristics, so confirm sizing with a pilot workload before committing to a SKU.

Reserved Instance vs Pay-As-You-Go for Fabric capacity: which is cheaper?

Compute break-even from live API prices: divide the amortized monthly reservation cost by the SKU's PAYG hourly cost to get break-even hours per month. If expected billed hours exceed that threshold, the Reserved Instance wins.

What is Autoscale Billing for Spark in Microsoft Fabric?

Autoscale Billing for Spark is an opt-in serverless model where Spark jobs are offloaded from the capacity and billed per Spark CU-hour on a separate meter. Bursting and smoothing are disabled for Spark, and the base SKU is sized only for non-Spark workloads.

Why does the Azure Retail Prices API return no rows for my Fabric meter filter?

Filters are case-sensitive exact matches, and invented meter names like 'OneLake Storage' or 'Spark Serverless' return zero rows. Fall back to listing all Microsoft Fabric meters for the region and select the correct tiered meter name.

Can I size a Fabric SKU from Synapse DWU using a formula?

No official DWU-to-CU conversion exists, so formula-based sizing is not actionable for committed SKUs. Pull actual DWU-hours from billing, run a pilot on a candidate F-SKU, measure CU-seconds in the Capacity Metrics app, and scale from that ratio.