When deploying Direct Lake semantic models via Deployment Pipelines, the Lakehouse source is incorrectly treated as an external SQL source. This results in broken lineage and prompts for credentials. Even when using deployment rules to update the SQL endpoint and database ID, the semantic model does not re-bind to the Lakehouse in the target environment. Expected behavior: Deployment Pipelines should natively support Direct Lake re-binding, maintaining proper lineage and avoiding unnecessary credential prompts. Current workaround: Using the semantic-link-labs library to rebind the semantic model programmatically post-deployment. ... View more

As a Fabric Administrator / Capacity admin in my organization, I have to work with many workspaces and not all of the assets (semantic modes, reports, dashboards) are built by an IT team. Some are built very jankly by novice users. Unfortunately the capacity is shared amongst all users and I have no way to limit the capacity consumption by workspace. That means that if a novice user deploys a bad data model to the capacity, all workspaces will essentially choke on itself and all users start complaining of slow dashboards and creating a negavtive impression of Power BI in the org.  I have been suggested by "Industry experts" to get another capacitity and assign all "development" workspaces to it, and vet datasets before deploying to production capacity. But again this only works if we have an army of IT resources (knowledgeable in Power BI people)  to vet ever single asset being deployed. But thats simply not the case. Additionally, we don't have unlimited budget to spend on Power BI / Fabric capacities unfortunately.  As such my idea is as follows: Somehwere in the admin portal / capacity settings, We add some kind of resource limitation setting by workspace. This way I can safely grant a novice developer a slice of the capacity without affecting the rest of my users on the tenant.  I could go on as to why this will benefit all of us in so many ways, but for the sake of brevity I'll leave it at that. If you'd like to understand the idea more please reach out to me.  ... View more

queryinsights currently supports only ASCII characters.   Specifically, both the command column in queryinsights.exec_requests_history and the last_run_command column in queryinsights.frequently_run_queries have non-ASCII characters converted to “? ”and the last_run_command column of queryinsights.frequently_run_queries are converted to “? ”.   In order to use queryinsights in auditing, it would be effective and useful to support non-ASCII characters.   We have been informed from a Microsoft support request that this limitation will be removed in a future feature addition. We therefore hope to have this resolved as soon as possible. ... View more

The customer requires key pair authentication for Snowflake as their organization no longer allows the use of usernames and passwords. Since Power BI does not support key pair authentication, this feature is needed in upcoming updates. ... View more

Story: As a report creator, I would like to be able to have all relevant data into a single report but be able to share specific report pages to distinct users.   Problem: Report creators are unable to apply user permissions to a report sheet/page. This forces users to create an entirely separate report instead of keeping everything consolidated into a single report. This increases time, complexity of management for distinct report stakeholders.   Workarounds: Report creators are able to create page navigation options and buttons to obscure and hide additional pages and apply those items to an RLS.   Implications with workaround: If the URL of reports are shared by other stakeholders or drill-down is used, you can access all the data and content being hidden from another page. This approach assumes the user will never obtain a link from another collaborative user. ... View more

It would be really useful to have security settings to show or hide individual report pages. Currently, audiences control access only at the report level, not per page. ... View more

We mirrored one of our Azure SQL databases to Fabric using the Mirrored Azure SQL database item. One of the table has about 500M records. In another workspace than the mirrored database, we have a Lakehouse in which we created shortcuts to the mirrored tables. We query the large table using pyspark with a date filter, to get all records of one or more dates into a dataframe. What we are noticing is that when filtering the table, a table scan occures over all 500M records. Although this takes a couple of minutes, we would like to avoid this table scan. Partitioning mirrored the table should help prevent this and should also decrease the amount of capacity CU's consumed to do this task.   I would love to have the possibility to configure partitions on a mirrored table. This should be possible in the initial phase when creating the mirrored database and also afterwards. When initializing the partitioned table, the partitions are immediately applied. When applying a partition afterwards, when a sync is already active, the table would probably have to resync into the new partition. ... View more

To address the challenges of organizing and presenting data due to the limitations of column and row headers, a potential solution would be to implement a new matrix visual with customizable controls, allowing report creators to adjust the dimensions of columns and rows, group them hierarchically, apply diverse styles, and use conditional formatting. ... View more

I would like to request full support connection from power bi service to azure databricks via Private Link without the need for deploying a vnet or on-premise gateway which could be potential bottlenecks.     ... View more

Please provide the ability in M query or DAX for a report to know it's context. That is; it knows its' Workspace ID or Report ID. If we had this, we can then enrich the report with context retrieved from Power BI REST API info (like the report's fabric Domain or other corporate governance info) ... View more

Problem: Currently, when a dataset is connected to a Lakehouse as a datasource in Power BI Fabric, it defaults to a cloud connection mapped to SSO. In embedded mode, reports built on these datasets fail due to a lack of identity, as they do not inherit authentication from the service principal. To resolve this, users must manually adjust the datasource settings via the Power BI service https://learn.microsoft.com/en-us/fabric/fundamentals/direct-lake-fixed-identity   Proposed Solution: Introduce a new REST API or extend the current API functionality to allow programmatic setting (or re-setting) of the connection to the service principal, including updating datasources. Benefits: Automates the process, reducing manual intervention Minimizes downtime for embedded reports Enhances developer experience and deployment efficiency Ensures consistency in authentication settings across different environments Impact: This feature will significantly improve the workflow for organizations embedding Power BI reports using service principals, ensuring seamless and automated datasource authentication. ... View more

Currently in Power BI, there is no built-in option to control or limit the number of selections a user can make in a slicer. While some custom visuals offer this feature, they behave unexpectedly — for example, the slicer automatically closes when the user exceeds the selection limit. Instead, it would be much better if Power BI provided a native option to limit slicer selections, such that users are prevented from selecting more than the allowed number of values, rather than allowing selection and auto-closing the slicer. This would greatly improve usability, especially in reports where we need to control filtering behavior strictly ... View more

Feature Request: Add Key Pair-Based Authentication for Snowflake Connector in Power BI Description: We request the addition of key pair-based authentication to the Snowflake Connector in Power BI to address current challenges with existing authentication methods. Microsoft Entra ID-based authentication requires setting up additional infrastructure and services solely for authentication, which increases complexity and operational overhead. Snowflake's recent enforcement of MFA for password-based authentication makes it difficult to configure and maintain connections on Power BI Gateway, as MFA is not compatible with automated workflows. Key pair-based authentication would provide a secure, lightweight, and efficient alternative, eliminating the need for extra infrastructure while simplifying the configuration process for Power BI Gateway. This feature would enhance usability and align with modern security practices. ... View more

With the debut of DAX Query View and TMDL Script View, there should be an API to retrieve these files.   This capability would enable organizations to automate testing and deployments by programmatically accessing DAX Queries and TMDL scripts.   For example, a Notebook could iterate over semantic models in a workspace, retrieve DAX Queries for testing, and execute those tests using semantic-link. If tests fail, teams could take corrective actions based on the results. ... View more

OPENROWSET function can read the files from Azure Data Lake storage or Azure Blob storage.  It would be useful to enable OPENROWSET to also read files form Fabric One Lake and access the files in Lakehouse /Files area. ... View more

To address the limitations of our current slicers, which are less advanced and potentially confusing compared to our competitors, a new and enhanced list and dropdown slicer should be designed with a wider range of customization options, incorporating industry best practices, and include innovative features like the ability to copy and paste lists. ... View more

The OneLake Catalogue in Fabric offers to filter by domain, but the same catalogue in Power BI Desktop does not offer the same filter. Can we have feature parity here? ... View more

Enable individual users / freelancers to study and practice fabric with a per-user license similar to Premium Per User in Power BI. F2 is a price point that is too high for them. Strong limitations in time and resources available would make sense, but there should be a clear monthly cap that cannot be like the current F2. ... View more

Make it possible to write SparkSQL without having a Default Lakehouse.   With 3- or 4- part naming, i.e. [workspace].[lakehouse].[schema].[table]    there should be no need to attach a Lakehouse in order to use SparkSQL.   Needing to attach a Lakehouse is annoying and adds extra complexity. ... View more