Microsoft Fabric Updates Blog

Capacity metrics in Microsoft Fabric

The team is excited to share an update on the Microsoft Fabric preview and the capacity metrics monitoring experiences in Fabric. Fabric builds on the SaaS foundation used by millions of Power BI Premium customers and extends its capabilities to unify a wide range of Microsoft Data experiences including Data Integration, Data Engineering, Real Time Analytics, ML and AI offerings and Business Intelligence with Power BI. In this post, we’ll focus on the universal compute capacities portion of the Fabric Intelligent data foundation and Fabric’s capacity metrics governance features that admins can use to monitor usage and make data-driven scale-up decisions to unlock the full potential of your organizations data.

Universal compute capacities

One of the most significant advantages of capacities is their ability to empower creators at scale across an organization with SaaS simplicity. Universal capacities reserve a specific amount of compute to drive Fabric experiences and enable an unlimited number of users to create or consume from Fabric experiences. Consumption of Fabric experiences is only bound by the amount of compute required to serve data to your users and can be easily increased with a SKU upgrade or on-demand through the use of autoscale.

A screenshot of a computer

Description automatically generated with low confidence

Fabric trial and licensing SKU’s

During this week’s Build event, we’re enabling the Fabric preview to let all customers try out the Fabric workload experiences on new trial capacities for 60 days or using their existing Power BI Premium P SKU capacities. For the first two months of the preview, only production Power BI workloads will drain capacity compute resources in purchased capacities.

The capacity and SKUs table lists the Microsoft Fabric SKUs. Each SKU is denoted by a size in Capacity Units (CU), which are used to measure the compute power available for that SKU. For the benefit of customers that are familiar with Power BI, the table also includes the equivalent Power BI SKUs and v-cores.

A screenshot of a computer

Description automatically generated with medium confidence

To learn more about starting a trial capacity read the trial documentation page.

Capacity metrics in Fabric overview

Capacity metrics in Fabric is a governance feature for admins to monitor the performance of workloads and their usage compared to purchased capacity. Capacity metrics unifies telemetry from all Fabric workload experiences into a single set of turnkey analytics.

In the 1st section below, capacity metrics uses a ribbon chart to show trends of usage across the different experiences hosted in Fabric. Data is aggregated daily, and you can also view the duration and type of operations along with user trends to measure adoption.

The 2nd section allows users to analyze the usage of universal compute capacity resources by the individual “Items” (previously known as artifacts) used across their Tenant. Usage is tracked by total Capacity Unit Seconds (CU(s)). Examples of Items include a Power BI report, a Dataflow or Synapse Notebook consuming data out of the open delta parquet format hosted in OneLake. Items are currently listed by: Item Name\Workload Experience Type\Workspace. The “Preview Status” field can be used to differentiate between production Power BI workloads and preview experience workloads that won’t drain capacity resources during the preview period.

A screenshot of a computer

Description automatically generated

The 3rd graph shows utilization of resources compared to the amount of throughput purchased for a given capacity. 100% of utilization represents the full throughput of a capacity SKU and is shared by all Fabric experiences. Larger capacity SKUs provide more throughput and scale for all workload experiences. When production workloads use 100% of capacity resources it may result in throttling. Preview workloads experiences can exceed 100% of capacity resources without throttling. Admins can use the percentage of consumption on their current SKU compared to SKU limits defined at Licensing and SKU documentation to easily plan for rollout of Fabric across the organization with confidence.

The Capacity Utilization view has been enhanced with Logarithmic scale for easy analysis of consumption during low utilization periods and has a new control to easily access Timepoint Drill via the explore button.

A screenshot of a graph

Description automatically generated with medium confidence

Capacity metrics in Fabric – Timepoint drill

While the main page of the capacity metrics experience in Fabric focuses on aggregate data, timepoint drill lets users zoom in to explore usage data with the full fidelity used by the SaaS capacity platform.

The amount of capacity provided by the given SKU per 30 second evaluation period is shown on the top section along with the breakdown of interactive (user driven) and background (service driven operations). In situations where user perceived duration of operations is less than the timepoint CU(s), users can see the advantage of platform “bursting” which runs operations across many back-end Microsoft managed capacity nodes to speed up execution.

Top use cases for this view include:

- Understanding what operations contributed to usage of capacity resources
- Determining what caused a capacity to autoscale
- Identification of a user who scheduled or ran an operation

A screenshot of a computer

Description automatically generated

Get started with Microsoft Fabric

Microsoft Fabric is currently in preview. Try out everything Fabric has to offer by signing up for the free trial—no credit card information required. Everyone who signs up gets a fixed Fabric trial capacity, which may be used for any feature or capability from integrating data to creating machine learning models. Existing Power BI Premium customers can simply turn on Fabric through the Power BI admin portal. After July 1, 2023, Fabric will be enabled for all Power BI tenants.

Once you’ve enabled your trial, you can access the latest version of Capacity Metrics experience in Fabric. Existing customers of Capacity Metrics app should use this link to install the latest version purpose-built for Fabric. By Fabric general availability (GA) P SKU customers will be required to use the latest version of the Capacity Metrics app. During the preview period, we’ll be listening to feedback from customers and incorporating it into platform evolution and updates to the Capacity Metrics app.