Roadmap tracker
Recent Microsoft Fabric roadmap updates for the community, with filters for product, release type, status, and search.
This page pulls roadmap data from the Fabric GPS JSON API (with an RSS fallback if needed). We cache and refresh this feed every 24 hours, so there can be a short delay between a Fabric blog/roadmap update being published and it appearing here. If Microsoft updates or edits an existing roadmap item, those changes are picked up on the next refresh cycle.
OneLake malware protection helps secure your data by automatically scanning files for malicious content when they're uploaded. It surfaces alerts and scan results for further action, providing built-in protection for enterprise data without requiring additional infrastructure.
**Business Events Data Preview Overview**Enable developers to inspect, validate, and troubleshoot Business Events when publishing them to downstream consumers. Data Preview provides real-time visibility into event payloads, schema mappings, and event metadata, reducing time-to-value and simplifying event-driven application development.**Key Capabilities*** Real-Time Event Inspection: Preview Business Event payloads as they are generated, allowing developers to validate event content before integrating downstream consumers.* Schema Visibility: View event schemas and payload structures side-by-side to ensure events conform to expected Business Event definitions.* Metadata Exploration: Inspect event properties, headers, correlation identifiers, timestamps, and other event metadata to better understand event context and flow.* Accelerated Troubleshooting: Quickly identify publishing issues, schema mismatches, missing attributes, and configuration errors without requiring external debugging tools.* Developer-Friendly Testing Experience: Validate Business Events during development and onboarding, enabling faster iteration and reducing time to first successful event-driven scenario.* Improved Confidence Before Consumption: Allow publishers to verify event quality and correctness before events are consumed by Eventstream, Activator, AI agents, or other downstream applications.* End-to-End Event Flow Validation: Enable developers to confirm that Business Events are generated as expected before investing effort in consumers, automation workflows, or integrations.
Eventstream now integrates with Fabric Spark, letting data engineers and scientists consume and process stream data in notebooks and Spark jobs. This unifies real-time intelligence with Spark in a single end-to-end experience.With this capability, you can:- Discover and consume Eventstream data directly from Fabric notebooks.- Process streaming data with Spark Structured Streaming jobs.- Build real-time analytics and AI applications on live event data.
We are adding support **dbt Fusion runtime** in dbt job in Microsoft Fabric as an execution environment, in addition to the existing dbt Core-based runtime.Customers can select the Fusion runtime when creating or configuring a dbt job to run their transformations on the next-generation dbt engine -- natively inside Fabric, with no local tooling or external orchestration to manage.
Fabric feature store is a centralized system that manages the data inputs, called features, that power machine learning models. It bridges raw data and ML models, giving teams one place to define, store, serve, and discover features instead of each team building its own pipelines.
Quickly identify when and where throttling is occurring across your Eventhouse system. This feature highlights recent throttling events, helping you diagnose performance bottlenecks and take corrective action before they impact users.
This feature enables customers to develop or reuse their own Kafka connectors, or open-source Kafka connectors, when no pre-built streaming connector is available.With this capability, you can:- Customize your own connector at your own pace.- Run with no infrastructure to manage.- Reuse open-source connectors directly.
Identify how applications like Activator and QuerySet contribute to Eventhouse activity. Analyze query volume and activity by application and identify unusual spikes to support informed capacity management.
You can now join data from OneLake shortcuts in update policies. Use this for dimension lookups, such as enriching ingested fact events with customer, device, or product attributes stored in OneLake shortcut data.The shortcut-backed external table must have Query Acceleration Policy enabled, and Hot duration must cover all data.
Branch workspace admin profile (aka. admin delegation) lets developers use the branch-out experience without needing create-workspace or assign-capacity permissions. Workspace admins keep full control through pre-configured guardrails while unblocking developer workflows. Capabilities: - Developers without create-workspace or assign-capacity permissions can perform branch-out seamlessly - Admins pre-configure profile to ensure guardrails for branch workspaces: developer role, capacity assignment and admin list.
List slicer with drop down mode and advanced search options.
Conditional formatting to be available for series in Power BI visuals. This includes lines, markers, and legends.
Data Clustering enables faster read performance by allowing users to specify columns for co-locating data on ingestion and perform file skipping on read.
Workspace identity auth support for Eventstream Event hub source. Customers can use workspace identity to do the authentication between Eventstream and their event hub source.
Fabric data engineers today write 80-100+ lines of boilerplate PySpark to ingest CSV and Parquet files into Delta tables - handling file discovery, schema inference, incremental refresh, schema drift, and error recovery manually. This feature introduces a simple Spark SQL DDL surface (CREATE MATERIALIZED LAKE VIEW ... FROM OneLake_Files OPTIONS (...)) that declaratively handles all of this. It includes automatic schema evolution (or strict fixed-schema mode), three refresh modes (append_only, mirror, full), built-in error handling, structured user telemetry on files and lakehouse-level DAG lineage.
Ability to clone warehouse across workspaces
**Release:** Public Preview | **Product area:** Fabric CI/CD**Deployment plan** is a new first-class Fabric workspace item that lets you **define, order, and automate** how your content is deployed across environments.Instead of relying on system-inferred lineage order, you author a plan on a **visual canvas** (with **YAML as the source of truth**) to:- **Declare the exact sequence** in which items deploy- **Add pre and post deployment steps** -- for example, running a Notebook to hydrate a Lakehouse, or a Data Pipeline before dependent items go liveDeployment plans **attach to the deployment operations available in Fabric**, including **Git** (branch-out, initial sync, update from Git), **Deployment Pipelines**, **Bulk Import**, and the **CI/CD library**, and are **honored at deploy time**.The result: every promotion follows your intended orchestration and produces a **working, reproducible solution**, not just a set of deployed definitions.
Ability to configure the data warehouse retention between 1 to 120 days;
Shortcuts in Fabric DW allow users to directly access and query data that is present in internal and external data sources, without loading them into Fabric DW. Users will have the capability to create Table Shortcuts via TSQL & UX
This feature is part of the concurrency control strategy for Fabric DW. File-level write-write conflict detection is a mechanism designed to prevent two concurrent transactions from modifying the same physical file in a data warehouse at the same time. More precise than table-level detection (which blocks any concurrent changes to the same table).What It Does?Scope: Operates at the file level (e.g., Parquet files in Fabric DW), rather than at the table or row level.Goal: Detect overlapping changes (updates, deletes) to the same file during concurrent transactions.Trigger: When a transaction tries to commit changes, the system checks if any newer manifests indicate modifications to the same file since the transaction started.Lays groundwork for even finer-grained detection (row-level) in later releases.
SQL analytics endpoints of Lakehouses will support large string and binary data using VARCHAR(MAX) and VARBINARY(MAX) types up to 16 MB.
For VNet data gateways, the fixed node setup can be inefficient during high-demand periods. To improve this, we're introducing auto-scaling, which dynamically adjusts the number of nodes based on demand.
We are introducing PostgreSQL as a new data destination for Dataflow Gen2, enabling customers to publish transformed data from Microsoft Fabric directly into PostgreSQL databases using Dataflow Gen2's low-code Power Query experience.This release expands Dataflow Gen2's destination ecosystem to better support operational, analytical, and application integration scenarios, giving customers a simple and flexible way to operationalize Fabric-based data transformations in PostgreSQL environments.Key benefits and scenarios:* Publish curated outputs from Dataflow Gen2 directly into PostgreSQL databases and tables* Support operational reporting, application integration, and downstream analytics scenarios using PostgreSQL as a serving layer* Enable teams to prepare, standardize, and reshape data in Fabric before delivering it to PostgreSQL-based applications, services, or analytics workloadsThe PostgreSQL data destination helps customers simplify data movement and reduce the need for custom pipelines by combining Fabric's scalable transformation capabilities with PostgreSQL's broad ecosystem compatibility and adoption.This release is part of our broader effort to make Dataflow Gen2 a flexible, low-code transformation layer across data platforms, with continued investments planned to expand destination support and simplify end-to-end data integration experiences in Fabric.
Private link allows customers to connect to their SQL DBs via a private endpoint from the customer's virtual network. Workspace level private links will give customers the flexibility to use private link connection to specific workspace(s) which have high security requirements and contain sensitive data while their other workspace can be open to public.
