Skip to main content

Announcing long-term usage and performance insights (Public Preview)

Headshot of article author Bhavik Merchant

Today, I am pleased to announce the availability of detailed historical activity information and analytical insights on Power BI operations. This new capability gives you full visibility into Power BI usage and performance so you can confidently manage your environment and comply with organizational policies. As a Workspace Administrator, you can now enable collection of Power BI user activities and metrics for any workspaces you own in near real-time. To assist you with data analysis, we provide you with a Power BI report template allowing you to:

  • Visualize historical data over long periods to observe usage trends, spot periods of unusual load, and identify performance degradation.
  • Isolate activities by time range, capacity, dataset, user, report and more.
  • Investigate internal operations within a query or dataset refresh to see where time is being spent.

This functionality addresses a common customer request to provide historical insights for workspace usage and performance data. This functionality takes advantage of your Azure environment, allowing data to be stored and retained for up to 2 years with Azure Log Analytics.

Historical insights are currently supported on Premium Gen1 and Gen2 workspaces hosted in Power BI Premium, Premium Per User (PPU) and Power BI Embedded capacities.

 

What insights can I get?

This first milestone in the insights roadmap exposes events from the Analysis Services engine sitting in the heart of Power BI. Analysis Services engine events are about Power BI datasets and are very useful in understanding engine load, query performance and refresh behavior. Our new capability allows logging of detailed engine level traces similar to those available today as Azure Analysis Services traces. However, we have added new details and made schema improvements such as friendlier names and strongly typed columns (e.g. integer type used for whole numbers). Additionally, we are providing quick start reporting templates to more easily facilitate your analysis for immediate value. For instance, the dataset activity report template is designed to help you answer common questions such as:

  • Which is the most popular dataset?
  • What operations generate the most load?
  • Who are the most active users?
  • What is my error rate?
  • Is query/refresh performance consistent?
  • Are there any periods of excessive activity?
  • What queries or refreshes are executed in a specific time period?
  • Which are the slowest operations?
  • What is contributing to query/refresh slowness?
  • What queries/refreshes spend the most time waiting for data from external sources?
  • What DQ queries is Power BI sending to external data sources?

 

The below example shows a workspace summary of engine activity, followed by a focus on query operations with a drill down into all the queries that ran on a specific day, from slowest to fastest.

 

The next example shows a query performance statistics overview. Here we can see which queries are using aggregations, examine the execution history of a particular query signature, and drill down to a single slow execution to explore which aggregations were used.

 

This specific report template enables you to more fully leverage the data model and DAX without requiring you to become an expert on Analysis Services log structure. It is the first of many built with Microsoft standard Fluent design guidelines, so you can expect a consistent look and feel as more analytics are released.

You can get this report in two ways:

– Download the .pbit template from GitHub, which you are free to customize

– Install the Log Analytics for Analysis Services Template App from AppSource (coming soon)

 

The report template is based on an Import dataset and does need to be periodically refreshed. The fastest way to analyze near real-time data is by querying the Log Analytics database directly, e.g., via the Azure portal. The example below demonstrates how to run queries in the Azure portal using the Kusto query language. Here, we quickly access a previously-run query to determine how many queries were executed per hour over the last 7 days, then view the results as a column chart.

 

 

How can I set up logging for my Workspace?

This initial release allows workspace administrators to configure logging for their Power BI workspaces and send data to a pre-configured Azure Log Analytics Workspace, a specialized database optimized for log data. First, the Power BI service administrator enables Log Analytics integration. Once Azure pre-requisites are met, workspace owners will be able to independently configure logging and access data.

A new Power BI Admin portal setting   under Azure Connections allows Power BI Service Administrators to enable Log Analytics capability for Workspace Admins:

With the tenant setting enabled, workspace administrators can configure logging for a v2 workspace by navigating to workspace settings and configuring the integration in the Azure Connections section, as shown below:

 

 

Once connected, data is sent continuously and is typically available in Log Analytics in about 5 minutes. The diagram below illustrates Azure Monitor at a high level, highlighting the path taken by Power BI.

As we enhance this feature to more broadly enable logging in Power BI, we plan to allow other administrators to configure logging independently for their specific scope of control.

We encourage you to check out the links below to learn more about our new activity tracking capability, and try it out for yourself.

 

Next Steps

– Read the feature documentation to learn more about configuration options and pre-requisites

– Create an Azure Log Analytics Workspace