The new calendar-based time intel is really good, but the fact that date-related columns (the ones that are not mapped to any calendar category, but still marked as date related using TMDL) are not removed when using DATEADD and SAMEPERIODLASTYEAR means that you need to write REMOVEFILTERS for doing time intel and that does not feel right at all.  Changing current behavior might be tricky but there could be a calendar level setting or yet another column category to make this happen. If I want a column to say as filter I will not mark it as date-related and that's it. If I mark it as date-related is because I don't want to deal with it when doing time inte. Please change that or give us a way around it. Writing REMOVEFILTERS on columns of the date table is very inconvenient and feels like a step back from legacy time intel.  ... View more

Need option for tenant‑wide UI view in Power BI Service that shows all subscriptions across the entire organization.   ... View more

Currently the personal bookmarks reference to text names of the columns in the reports hence when we make a change in the field name the personal bookmarks break. With the addition of PBIP format column names should be able to reference IDs hence we would be able to change the column name but keeping the column ID the same (this is only a suggested mechanism that comes to mind). This would provide flexibility for the development process and adaptibility to the business terminology changes in the future. ... View more

Currently, Power BI subscriptions can only be managed at an individual report or user level. There is no centralized option for administrators to disable subscriptions at a capacity level or for specific workspaces.  Need a feature to turn off subscriptions for an entire capacity or selected workspaces would greatly improve governance, performance control, and operational efficiency. ... View more

In Microsoft Fabric / Power BI, when multiple reports are connected to the same semantic model, Git integration does not preserve or allow control over the default report associated with that semantic model after syncing to a new workspace. This results in unexpected behavior when downloading, modifying, and republishing semantic models—especially after pulling artifacts into a new workspace using Git integration.   Scenario / Reproduction Steps A report is created in Power BI Desktop and published to the Power BI Service. This creates Report 02 test and Semantic Model 02 test. Later, another report is created directly in the Power BI Service using the same semantic model. This creates Report 01 test, also linked to the same semantic model. Both Report 01 test and Report 02 test are linked to the same semantic model in the original workspace, and everything works as expected. Git integration is enabled for the workspace, and all artifacts are pushed to a Git repository. The same artifacts are then pulled into a new workspace using Git integration. Observed Behavior (Issue) In the new workspace, the semantic model automatically selects Report 01 test as the default report. The selection appears to be based on alphabetical order, not on the original or intended default report. When the semantic model is downloaded as a PBIX, modified, and republished: The changes are reflected in Report 01 test instead of the 02 test Report. This issue does not occur in the original workspace and happens only after Git pull into a new workspace. Impact Creates confusion for Fabric admins and developers managing multiple reports per semantic model. Causes unexpected report updates, leading to potential data integrity and governance concerns. Makes Git‑based CI/CD workflows unreliable when: Multiple reports share a single semantic model Default report selection matters for downstream operations. Git integration should provide a way to explicitly define or preserve the default report associated with a semantic model, instead of automatically selecting one based on alphabetical order.   Feature Request Add an option in Git integration / Fabric workspace settings to: Explicitly select or configure the default report for a semantic model or Preserve the original default report association when syncing across workspaces. ... View more

In need to programmatically retrieve the status of all gateways in a cluster, not just the primary. Specifically, require status per gateway, detection of states like 'Partially offline', and alerting when any gateway becomes unavailable. Currently, APIs seem to return only the primary gateway status, with no visibility into secondary nodes and also alerting options, such as triggering alerts if a threshold of refresh failures is exceeded, for example, 20% failures within the last hour to ensure high availability and monitoring. ... View more

When a parameter is created with numeric range and data type decimal with maximum value 5000 and increment of 0.1 and used in a slicer, the input numbers rounds off to nearest whole number. Example: when user enters 2 in the slicer, then the input is taken as 0 and if 3 is entered then rounds to 5 which is the nearest whole number. When a parameter reaches the limit of 1000 as per document, additional values cannot be loaded, which restricts its usability in real‑world enterprise scenarios. Many organizations rely on parameters to dynamically filter large datasets, such as user lists, customer IDs, product catalogs, or report‑level security mappings. In such cases, the 1,000‑value ceiling becomes a hard blocker, forcing customers to redesign models, split reports, or abandon parameter‑based logic entirely. Increasing this limit would significantly improve scalability and reduce unnecessary complexity in report and model design..   Conclusion: Increasing the limit from 1000 to higher value is proposed to meet business requirements. Ref doc: Use Parameters to Visualize Variables - Power BI | Microsoft Learn  ... View more

Input Slicer   Currently, users are able to choose from a list of filter operators in the dropdown.   It would be helpful to be able to restrict which filter operators are available in the drop down for end-users to choose from.   Case 1: Limit to a single filter operator - e.g. Is Any Case 2: Limit the filter operator to 2 operators - e.g. Is Any, Contains Any etc... ... View more

Working scenario: In an existing report the matrix Bold formatting worked as expected. After selecting/unselecting bold in row headers works as expected. When opened a new page and created a new matrix visual selecting bold/unbold in row header and row sub totals work as expected. However when created a new report with the same model and exact same columns the matrix visual automatically bolds the first column. When additional column is added it appears has bold but only the last column works as expected. If row header is toggled between bold or unbold, it just bolds/unbolds the last column . If additional column is added and select row subtotals it bolds the total along with the first column. We received an update from Engineering team informing that it is by design as of now "The subtotal row headers are always rendered with the bold font family regardless of the 'Row Headers' bold toggle. The toggle doesn't control subtotal row header font which has its own forced bold styling". The working and non working has a different theme which explains why the old/new PBIX are behaving the way they are. This needs to be changed and bold formatting should work as expected in the existing report. ... View more

Distinct counts are the bane of our existence in big data.  Distinct counts over DirectQuery/DirectLake are not fast or efficient enough, and are very expensive from a compute and cost standpoint.  We need a re-aggregable approximate distinct count in Import Mode using HyperLogLog (HLL).  The idea would be to enable the creation/ingestion of a HLL sketch at the same dimensional grain as other user-defined aggregations so that the sketches can be dynamically merged/aggregated to produce an approximate distinct count.     Consider the following example where we need a distinct count of orders across channels and product groups.  Today, we would need to cache 4 unique intersections of data (1. Overall, 2. By Channel, 3. By Product Group, 4. By Channel & Product Group).  Not only is this compute intensive, but it forces us to create really terrible DAX to detect the user's reporting context and return the correct count.  In addition, it still doesn't allow for correct counts when users arbitrarily include/exclude members of a desired grouping (i.e. all product groups except "PRODUCT GROUP 02").  In an ideal world for this example, we would compute an HLL sketch at the Channel and Product Group level (image below), and either have a new DAX function like "HLLMERGE( <sketch> )" or enhance the APPROXIMATEDISTINCTCOUNT function to accept a column of sketchs, and this would dynamically merge the sketches in context to return the correct approximate distinct count. This would be a huge unlock to be able to dynamically reaggregate approximate distinct counts, as they are very difficult to manage today, and Power BI is well-positioned to do this.  It would be a powerhouse feature for big data users. ... View more

Hi, during the design of a task flow in a workspace it may happen that a task is accidentally deleted, but it doesn't it exist a manner to restore it: for a such case an undo feature would be very useful to restore the deleted task. Thanks ... View more

Enable users to configure the refresh frequency of shortcut transformations so they can balance data freshness and cost. This would allow users to choose between near real-time usage when up-to-date data is required and lower-frequency refreshes when they want to reduce cost and ingest data more easily. ... View more

When visiting the monitor page, not all refresh events appear. Monitor page does not show all semantic model refresh failures. The failed refresh events shows in refresh history of semantic models. Even the details are available in onelake catalog when searched for the semantic model -> monitor -> we could see the failures events but in monitor hub the details are not available. The Engineering team confirmed it by design but however the monitor hub should provide all the details so that it provides a centralized view of job execution health, progress, and outcomes to quickly identify problems and take action. ... View more

Idea Provide an option in Microsoft Fabric Oracle database mirroring to skip or exclude specific columns during replication, especially Oracle system-generated hidden columns (SYS_NC*). Problem Oracle mirroring fails during the initial load when tables contain hidden SYS_NC columns, raising ORA-00904. Currently, users must modify source Oracle databases (for example, drop extended statistics), which is not always acceptable in production environments. Benefit Enables mirroring without source-side schema or statistics changes Improves compatibility with enterprise Oracle databases Reduces onboarding blockers and support escalations Suggested Solution Allow column-level exclusion (or automatic exclusion of hidden columns) during initial load and ongoing replication. ... View more

Scheduling Hub   It would be great to to be able to see which items in your fabric workspace (and capacity) have schedules and if they are on or off. A bonus feature would allow you to enable and disable the schedules right from the list view (editing could still open the item in the editor and open the scheduling pane).   Being able to manage schedules from a central location is helpful in 'emergency' scenarios where you need to turn off all the pipelines and notebooks for a workspace. It could also be useful to ensure developers are not mismanaging schedules in Dev or Feature Branch workspaces (where we might typically not want them running). A whole set of useful columns are possible: Item Name  <click on the name to go to the item!> Item Type Workspace Schedule Type Next Scheduled Run Last Scheduled Run   Icons.. Edit Pencil, Enable/Disable, Trash Can The possibilities are endless!   ... View more

I would like to request that Dark Mode be added to the Fabric UI on the web. It can be quite jarring for those of us who use Dark Mode regularly to be blasted with the current UI. I saw there is a Dark Mode already in the Cosmos DB panel. Bringing that to the entire UI would make so many of us very happy and our eyeballs would thank you. ... View more

The current card visual forces users to overlap elements or waste copious amounts of time creating custom visuals. The new card feature should give users the ability to create multiple cards in a single container and provide a greater level of customization. ... View more

It would be beneficial to incorporate features from Pivot tables that allow for the expansion and collapse of columns and hierarchical column groups within tabular visuals. This would not only solve the current limitations of matrices but also provide report creators with the flexibility to hide and show rows and columns, saving these settings for future use, thus eliminating the need to scroll through irrelevant data. ... View more

Hi Community, As someone who has taken advantage of the .pbip folder structure and git integration with PBI I am very pleased with the functionality but I have noticed a few items that would be an improvement over the current setup. For reference, I am utilizing a general workflow of staging branch linked to dev workspace, main linked to prod workspace, devs create feature branches from staging and then PRs are opened into staging from feature and then PRs off of release branch from staging is opened with main. On each successful commit to staging or main a github action runs that syncs the branches contents with the online workspace. Now this brings me to my point, there are times when devs would like to make a minor change directly within the workspace. I want to get this changes tracked and incorporated in git; however, there is currently only the option for commitng directly to the linked branch or creating and commiting to new branch (and then you could follow the stated dev cycle I outlined above assuming you create a new branch) BUT it would be greatly appreciated if you could open a PR from what is in the online workspace with the linked branch because my staging and main branches are protected and no commits can be made directly to then without review. I am curious if anyone else has a similair experience or if this would be helpful for them? Please a few screenshots below for reference.    Any interaction on this post is greatly appreciated.   ... View more