Fabric operations

Each experience within Microsoft Fabric supports unique operations. An operation's consumption rate is what converts the usage of the experience's raw metrics into Compute Units (CU).

The Microsoft Fabric Capacity Metrics app's compute page provides an overview of your capacity's performance and lists Fabric operations that consume compute resources.

This article lists these operations by experience, and explains how they consume resources within Fabric.

Interactive and background operations

Microsoft Fabric divides operations into two types, interactive and background. This article lists these operations and explains the difference between them.

Interactive operations

On-demand requests and operations that can be triggered by user interactions with the UI, such as data model queries generated by report visuals, are classified as interactive operations. They're usually triggered by user interactions with the UI. For example, an interactive operation is triggered when a user opens a report or clicks on a slicer in a Power BI report. Interactive operations can also be triggered without interacting with the UI, for example when using SQL Server Management Studio (SSMS) or a custom application to run a DAX query.

Background operations

Longer running operations such as semantic model or dataflow refreshes are classified as background operations. They can be triggered manually by a user, or automatically without user interaction. Background operations include scheduled refreshes, interactive refreshes, REST-based refreshes and XMLA-based refresh operations. Users aren't expected to wait for these operations to finish. Instead, they might come back later to check the status of the operations.

How to read this document

Each experience has a table that lists its operations, with the following columns:

When more details regarding the consumption rate are available, a link to the document with this information is provided.

Fabric operations by experience

This section is divided into Fabric experience. Each experience had a table that lists its operations.

Important

Consumption rates are subject to change at any time. Microsoft will use reasonable efforts to provide notice via email or through in-product notification. Changes shall be effective on the date stated in Microsoft's Release Notes or Microsoft Fabric blog. If any change to a Microsoft Fabric Workload Consumption Rate materially increases the Capacity Units (CU) required to use a particular workload, customers might use the cancellation options available for the chosen payment method.

Copilot in Fabric

Copilot operations are listed in this table. You can find the consumption rates for Copilot in Copilot consumption.

Operation Description Item Azure billing meter Type
Copilot in Fabric Compute cost associated with input prompts and output completion Multiple Copilot in Fabric CU Background

Data Factory

The Data Factory experience contains operations for Dataflows Gen2 and Pipelines.

Dataflows Gen2

You can find the consumption rates for Dataflows Gen2 in Dataflow Gen2 pricing for Data Factory in Microsoft Fabric.

Operation Description Item Azure billing meter Type
Dataflow Gen2 Refresh Compute cost associated with dataflow Gen2 refresh operation Dataflow Gen2 Dataflows Standard Compute Capacity Usage CU Background
High Scale Dataflow Compute - SQL Endpoint Query Usage related to the dataflow Gen2 staging warehouse SQL endpoint Warehouse High Scale Dataflow Compute Capacity Usage CU Background

Pipelines

You can find the consumption rates for Pipelines in Data pipelines pricing for Data Factory in Microsoft Fabric.

Operation Description Item Azure billing meter Type
DataMovement The amount of time used by the copy activity in a Data Factory pipeline divided by the number of data integration units Pipeline Data Movement Capacity Usage CU Background
ActivityRun A Data Factory data pipeline activity execution Pipeline Data Orchestration Capacity Usage CU Background

Data Warehouse

One Synapse Data Warehouse core (unit of compute for Data Warehouse) is equivalent to two Fabric Capacity Units (CUs).

Operation Description Item Azure billing meter Type
Warehouse Query Compute charge for all user generated and system generated T-SQL statements within a Warehouse Warehouse Data Warehouse Capacity Usage CU Background
SQL Endpoint Query Compute charge for all user generated and system generated T-SQL statements within a SQL Endpoint Warehouse Data Warehouse Capacity Usage CU Background

OneLake

One Lake compute operations represent the transactions performed on One Lake items. The consumption rate for each operation varies depending on its type. For more details, refer to One Lake consumption.

Operation Description Item Azure Billing Meter Type
OneLake Read via Redirect OneLake Read via Redirect (Multiple) OneLake Read Operations Capacity Usage CU Background
OneLake Read via Proxy OneLake Read via Proxy (Multiple) OneLake Read Operations via API Capacity Usage CU Background
OneLake Write via Redirect OneLake Write via Redirect (Multiple) OneLake Write Operations Capacity Usage CU Background
OneLake Write via Proxy OneLake Write via Proxy (Multiple) OneLake Write Operations via API Capacity Usage CU Background
OneLake Iterative Write via Redirect OneLake Iterative Write via Redirect (Multiple) OneLake Iterative Write Operations Background
OneLake Iterative Read via Redirect OneLake Iterative Read via Redirect (Multiple) OneLake Iterative Read Operations Capacity Usage CU Background
OneLake Other Operations OneLake Other Operations (Multiple) OneLake Other Operations Capacity Usage CU Background
OneLake Other Operations via Redirect OneLake Other Operations via Redirect (Multiple) OneLake Other Operations via API Capacity Usage CU Background
OneLake Iterative Write via Proxy OneLake Iterative Write via Proxy (Multiple) OneLake Iterative Write Operations via API Capacity Usage CU Background
OneLake Iterative Read via Proxy OneLake Iterative Read via Proxy (Multiple) OneLake Iterative Read Operations via API Capacity Usage CU Background
OneLake BCDR Read via Proxy OneLake BCDR Read via Proxy (Multiple) OneLake BCDR Read Operations via API Capacity Usage CU Background
OneLake BCDR Write via Proxy OneLake BCDR Write via Proxy (Multiple) OneLake BCDR Write Operations via API Capacity Usage CU Background
OneLake BCDR Read via Redirect OneLake BCDR Read via Redirect (Multiple) OneLake BCDR Read Operations Capacity Usage CU Background
OneLake BCDR Write via Redirect OneLake BCDR Write via Redirect (Multiple) OneLake BCDR Write Operations Capacity Usage CU Background
OneLake BCDR Iterative Read via Proxy OneLake BCDR Iterative Read via Proxy (Multiple) OneLake BCDR Iterative Read Operations via API Capacity Usage CU Background
OneLake BCDR Iterative Read via Redirect OneLake BCDR Iterative Read via Redirect (Multiple) OneLake BCDR Iterative Read Operations Capacity Usage CU Background
OneLake BCDR Iterative Write via Proxy OneLake BCDR Iterative Write via Proxy (Multiple) OneLake BCDR Iterative Write Operations via API Capacity Usage CU Background
OneLake BCDR Iterative Write via Redirect OneLake BCDR Iterative Write via Redirect (Multiple) OneLake BCDR Iterative Write Operations Capacity Usage CU Background
OneLake BCDR Other Operations OneLake BCDR Other Operations (Multiple) OneLake BCDR Other Operations Capacity Usage CU Background
OneLake BCDR Other Operations Via Redirect OneLake BCDR Other Operations Via Redirect (Multiple) OneLake BCDR Other Operations via API Capacity Usage CU Background

Power BI

The usage for each operation is reported in CU processing time in seconds. Eight CUs are equivalent to one Power BI v-core.

Note

The term Semantic model replaces the term dataset. You may still see the old term in the UI until it is completely replaced.

Operation Description Item Azure billing meter Type
Artificial intelligence (AI) AI function evaluation AI Power BI Capacity Usage CU Background
Background query Queries for refreshing tiles and creating report snapshots Semantic model Power BI Capacity Usage CU Background
Dataflow DirectQuery Connect directly to a dataflow without the need to import the data into a semantic model Dataflow Gen1 Power BI Capacity Usage CU Interactive
Dataflow refresh An on-demand or scheduled background dataflow refresh, performed by the service or with REST APIs. Dataflow Gen1 Power BI Capacity Usage CU Background
Semantic model on-demand refresh A background semantic model refresh initiated by the user, using the service, REST APIs, or public XMLA endpoints Semantic model Power BI Capacity Usage CU Background
Semantic model scheduled refresh A scheduled background semantic model refresh, performed by the service, REST APIs, or public XMLA endpoints Semantic model Power BI Capacity Usage CU Background
Full report email subscription A PDF or PowerPoint copy of an entire Power BI report, attached to an email subscription Report Power BI Capacity Usage CU Background
Interactive query Queries initiated by an on-demand data request from a user. For example, loading a model when opening a report, or user interaction with a report Semantic model Power BI Capacity Usage CU Interactive
PublicApiExport A Power BI report exported with the export report to file REST API Report Power BI Capacity Usage CU Background
Render A Power BI paginated report exported with the export paginated report to file REST API Paginated report Power BI Capacity Usage CU Background
Render A Power BI paginated report viewed in Power BI service Paginated report Power BI Capacity Usage CU Interactive
Web modeling read A data model read operation in the semantic model web modeling user experience Semantic model Power BI Capacity Usage CU Interactive
Web modeling write A data model write operation in the semantic model web modeling user experience Semantic model Power BI Capacity Usage CU Interactive
XMLA read XMLA read operations initiated by the user, for queries and discoveries Semantic model Power BI Capacity Usage CU Interactive
XMLA write A background XMLA write operation that changes the model Semantic model Power BI Capacity Usage CU Background

Real-Time Intelligence

The Real-Time Intelligence experience contains operations for Event streams and KQL Database and KQL Queryset.

Event streams

You can find the consumption rates for Event streams in Monitor capacity consumption for Microsoft Fabric event streams.

Operation Description Item Azure billing meter Type
Eventstream Per Hour Ingestion or processing for Event Stream Event Stream eventstream Capacity Usage CU Background
Eventstream Data Traffic per GB Data Ingress and Egress Event Stream eventstream Data Traffic per GB Capacity Usage CU Background
Eventstream Processor Per Hour ASA Processing Event Stream eventstreams Processor Capacity Usage CU Background

KQL Database and KQL Queryset

You can find the consumption rates for KQL Database in KQL Database consumption.

Operation Description Item Azure billing meter Type
KustoUpTime Measure of the time that the KQL database is Active KQL Database or KQL Queryset KQL Database Capacity Usage CU Interactive

Spark

Two Spark VCores (a unit of computing power for Spark) equals one capacity unit (CU). To understand how Spark operations consume CUs, refer to spark pools.

Operation Description Item Azure billing meter Type
Lakehouse operations Users preview table in the Lakehouse explorer Lakehouse Spark Memory Optimized Capacity Usage CU Background
Lakehouse table load Users load delta table in the Lakehouse explorer Lakehouse Spark Memory Optimized Capacity Usage CU Background
Notebook run Synapse Notebook runs manually by users Synapse Notebook Spark Memory Optimized Capacity Usage CU Background
Notebook HC run Synapse Notebook runs under the high concurrency Spark session Synapse Notebook Spark Memory Optimized Capacity Usage CU Background
Notebook scheduled run Synapse Notebook runs triggered by notebook scheduled events Synapse Notebook Spark Memory Optimized Capacity Usage CU Background
Notebook pipeline run Synapse Notebook runs triggered by pipeline Synapse Notebook Spark Memory Optimized Capacity Usage CU Background
Notebook VS Code run Synapse Notebook runs in VS Code. Synapse Notebook Spark Memory Optimized Capacity Usage CU Background
Spark job run Spark batch job runs initiated by user submission Spark Job Definition Spark Memory Optimized Capacity Usage CU Background
Spark job scheduled run Synapse batch job runs triggered by notebook scheduled events Spark Job Definition Spark Memory Optimized Capacity Usage CU Background
Spark job pipeline run Synapse batch job runs triggered by pipeline Spark Job Definition Spark Memory Optimized Capacity Usage CU Background
Spark job VS Code run Synapse Spark job definition submitted from VS Code Spark Job Definition Spark Memory Optimized Capacity Usage CU Background