Welcome to the April 2026 Microsoft Fabric update!
This month’s update brings a broad set of new capabilities across Microsoft Fabric, spanning the platform experience, Data Engineering, Data Science, Data Warehouse, and Real-Time Intelligence. Read on to learn about improvements to the Fabric experience, deeper VS Code integration, enhanced notebook resiliency, expanded machine learning and governance features, and new real-time data processing capabilities.
Contents
- Events and Announcements
- Ready to become a Microsoft Certified SQL AI Developer?
- A new Data Days event is coming soon!
- Watch the FabCon SQLCon recap series
- FabricPlatform
- Tabbed multitasking and object explorer update (Generally Available)
- Create AI Auto-Description for Semantic Models (Preview)
- Deprecating the Built-In Netezza ODBC Driver
- Data Engineering
- Retry policy support in notebook
- Manage Fabric workspace item inside VS Code
- Manage Environment item inside VS Code
- Set default Lakehouse of notebook VS Code
- Maven support in Fabric Environment (Preview)
- Data Science
- OAP support for ML experiments and models
- Manage ML experiments and models across workspaces and platforms
- Semantic Link (SemPy) 0.14.0: admin APIs and more
- Data Warehouse
- ALTER TABLE in explicit user transaction (Generally Available)
- COPY INTO support for JSONL files
- Real-TimeIntelligence
- Stream Mirrored Database change feeds into Eventstreams (Preview)
- Eventstream Observability with Workspace Monitoring (Preview)
- Eventhouse remote MCP (Preview)
- Eventhouse OneLake now supports schema evolution
- Configure and manage Activator rules directly in Eventstream (Preview)
- Custom CA and mTLS support in Eventstream connectors (Preview)
Events and Announcements
Ready to become a Microsoft Certified SQL AI Developer?
Introducing a new series to help SQL developers prepare for the new DP‑800 exam, Microsoft’s new certification for building AI‑enabled database solutions across SQL Server, Azure SQL, and SQL databases in Microsoft Fabric. The series breaks the exam down into clear, practical sessions that map directly to real-world database development.
By the end, you’ll know what to study, how the skills fit together, and how to apply them in real projects, not just on test day. The sessions are available live and on demand, in English, Portuguese and Spanish.
Sessions start the first week of May, register now.
A new Data Days event is coming soon!
In 2026, we’re going bigger than ever. Data Days will now expand beyond Fabric and Power BI and include SQL and AI. We are planning over 100 live sessions, more than 5 contests and challenges, and dozens of study groups and learning opportunities. You aren’t going to want to miss this event.
It all starts on June 15, 2026. Pre-register to get updates on Data Days activities, offers and more.
Watch the FabCon SQLCon recap series
Missed FabCon SQLCon this year, or couldn’t attend every session? This is your fast track to staying in the loop. In this series, we break down the biggest announcements from the event into clear insights and standout demos you can use right away. If it mattered at FabCon SQLCon, you’ll find it here.
Join us, get inspired, and stay connected to what’s happening across the Fabric community. Check out the The FabCon & SQLCon Recap Series now.
Fabric Platform
Tabbed multitasking and object explorer update (Generally Available)
After months of preview feedback, iteration, and performance tuning, both capabilities are now fully supported and available to all Fabric users.
Figure: Tabbed multitasking and object explorer in Fabric.
Improvements shipped since preview include:
- Enhanced tab management with right‑click options to customize tab layout.
- Ability to open tabs in a new browser window via right‑click.
- Resizable object explorer pane width for better navigation.
- Support for opening monitoring jobs as tabs.
- Accessibility improvements.
To learn more about all capabilities, refer to the multitask with tabbed navigation documentation.
Create AI Auto-Description for Semantic Models (Preview)
Understanding what your data contains shouldn’t require deep exploration. With the new Auto-Description capability for semantic models, you can use AI to instantly generate clear, meaningful descriptions that help others quickly understand your data.
Auto-Description uses the model’s metadata and structure to create a summary that explains what the semantic model includes and how it can be used. This makes it easier for teams to discover relevant data, build on existing work, and confidently choose the right assets.
Model owners and contributors can generate descriptions directly from the semantic model details page using Copilot. Once generated, the description can be reviewed, edited directly on the same page, and applied, so you stay in control and can ensure it reflects the right business context. You can also regenerate it at any time to refine or improve the result.
Figure: Generate and refine AI-powered descriptions for your semantic models directly from the details page.
Learn more about the OneLake catalog in the OneLake catalog overview documentation.
Deprecating the Built-In Netezza ODBC Driver
The IBM Netezza ODBC driver has been Generally Available for several weeks, and we are now starting the deprecation of the previously built-in ODBC driver to this new version.
Customers do not need to install the new connector; you may reuse your existing connector but will need to install the new ODBC driver. We encourage customers to do this as soon as possible to ensure a smooth transition. 
Figure: Image of IBM Netezza connector selection screen within Fabric.
Learn more with the IBM Netezza database documentation.
Data Engineering
Retry policy support in notebook
Fabric Notebook now supports a retry policy, allowing notebook jobs to automatically restart following system errors. For instance, if a Spark cluster is recycled, the notebook job will resume on the new cluster without requiring manual intervention.
To opt-in the retry policy, you can include the “retriableOptions “section in the %%configure command:
%%configure
{
"retriableOptions":
{
"maxAttempt": 3 # set the max time of the attempt
"enabled": true
}
}
This setup restarts the notebook up to three times after a job failure, with about 120 seconds between attempts. If it fails after three attempts, the execution will be canceled.
The retry policy only supports the notebook execution triggered by public API, Data Integration pipeline or the Eventstream. The interactive run is not supported in this release.
To learn more, refer to the Retry support inside Notebook documentation.
Manage Fabric workspace item inside VS Code
The Fabric Data Engineering VS Code extension lets you manage multiple fabric workspaces and their Data Engineering items within a single VS Code window. When the notebook item is open in this view, users can directly update the Notebook code inside VS Code, and the changes are automatically synced back to the remote workspace.
Select Open a Remote Windows inside VS Code and select Open Fabric Data Engineering Workspace, this should activate the VS Code explore view in the remote view.
Figure: Open Remote Window.
Figure: Open Fabric Workspace.
Select Manage Fabric Workspace to add the target workspace into this view or exclude any existing ones from this view.
Figure: Manage Fabric workspace.
To learn more, refer to the manage fabric workspace inside VS Code documentation.
Manage Environment item inside VS Code
The Fabric Data Engineering VS Code extension now lets you view and edit Environment items directly in VS Code. The Environment items are listed under the Environment node after the workspace is open under the VFS mode. To inspect the Environments detail, right-click the environment item and select Inspect Environment. This opens the environment details as a read-only .yml file in the VS Code editor, these are the settings already been published.
Figure: Select Inspect Environment.
To change the environment configuration, right-click the environment and select Edit Environment. This opens an editable .yml file in the VS Code editor
Figure: Edit Environment.
To learn more, refer to the manage Spark Environment inside VS Code documentation.
Set default Lakehouse of notebook VS Code
VS Code now lets you update the default Lakehouse for a notebook. Add or remove Lakehouses linked to the notebook—changes automatically sync back to the remote workspace. To view the linked Lakehouses, expand Dependencies, then Lakehouses. To set the default Lakehouse, right-click the Lakehouse and select Set as Default Lakehouse.
Figure: Set Default Lakehouse.
To learn more, refer to the develop notebook in VS Code documentation.
Maven support in Fabric Environment (Preview)
Fabric Environment supports adding libraries from Maven repositories. This lets Scala and Java developers manage libraries and transitive dependencies using standard Maven configurations. With this update, you can upload and manage pom.xml files directly in an Environment, eliminating the need to download JARs from Maven and upload them.
Figure: Import and export pom.xml.
To learn more, refer to the Manage libraries in Fabric environments documentation.
Data Science
OAP support for ML experiments and models
Machine Learning Experiments and Machine Learning Models can now be created and managed in workspaces with outbound access protection (OAP) enabled. Previously, these item types were blocked in protected workspaces. With this update, data science teams in security-sensitive environments can track experiments, log metrics, and register models without compromising their workspace security posture.
Figure: Animated gif - outbound access protection support for ml experiment and model.
To learn more, refer to the Workspace outbound access protection for Data Science documentation.
Manage ML experiments and models across workspaces and platforms
Fabric supports cross-workspace MLflow logging for experiments and models, enabling end-to-end MLOps workflows from development to test to production using standard MLflow APIs.
Additionally, you can bring existing ML assets from Azure Databricks, Azure Machine Learning, local environments, or any platform that supports MLflow directly into Fabric, making it easy to consolidate your ML artifacts without rebuilding training pipelines. Cross-workspace logging is also supported in workspaces with outbound access protection (OAP) enabled.
To get started, install the synapseml-mlflow package, set the tracking URI to your target workspace, and use the MLflow commands you’re familiar with.
Figure: Animated gif - Log ml experiment and model from databricks to Fabric.
To learn more, refer to Manage MLflow models across workspaces and platforms.
Semantic Link (SemPy) 0.14.0: admin APIs and more
SemPy 0.14.0 introduces the new sempy.fabric.admin module with 75 admin APIs for end-to-end tenant management, covering workspaces, capacities, domains, tenant settings, reports, users, and more—all directly from Python.
This release also enables deploying semantic models across workspaces with automatic remapping of Direct Lake connections to new lakehouses or warehouses, unlocking real CI/CD patterns for BI. Additional highlights include programmatic extraction and updating of Power BI report layouts via get_report_json and update_report_from_reportjson, Delta OPTIMIZE (compaction) on lakehouse tables from Python, and fine-grained control over long-running operations with the new LroConfig.
To learn more, refer to the full release notes.
Data Warehouse
ALTER TABLE in explicit user transaction (Generally Available)
Explicit user transaction enables supported ALTER TABLE operations to run within a user‑defined transaction (BEGIN TRAN … COMMIT) in Fabric Data Warehouse. Previously, ALTER TABLE could not participate in explicit user transactions even though other DDL and DML operations could, which meant schema migrations that combined ALTER TABLE with other operations could not be executed atomically in a single transaction.
This added deployment complexity and increased the risk of failures that left schemas partially updated. By bringing true atomicity to schema changes, this feature improves correctness, reduces deployment complexity, and makes enterprise schema migrations safer and more reliable.
For more information refer to the AlterTableInsideExplicitTransactions blog post.
COPY INTO support for JSONL files
COPY INTO support for JSONL documents enables customers to ingest newline‑delimited JSON (JSONL) files directly into Microsoft Fabric Data Warehouse tables using the existing COPY INTO surface area. Previously, while OPENROWSET supported JSONL, customers could not use COPY INTO for JSONL ingestion, forcing them to adopt alternative or more fragmented ingestion patterns.
With this capability, JSONL is now a first‑class FILE_TYPE option, allowing customers to load semi‑structured data at scale using the same trusted, performant, and observable ingestion workflow they already use for CSV and Parquet.
This closes a key gap for companies working with event streams, logs, and application exports commonly emitted in JSONL format, improving ingestion consistency, lowering operational complexity, and aligning COPY INTO with real‑world data formats used across modern data pipelines.
For more details, refer to the product documentation on COPY INTO (Transact-SQL).
Real-Time Intelligence
Stream Mirrored Database change feeds into Eventstreams (Preview)
Customers using Extended Capabilities in Mirroring to capture Delta Change Data Feed (CDF) can now stream those row-level changes directly into Fabric Eventstreams for low-latency, event-driven processing.
The new Mirrored Database Change Feed connector provides a fully managed path from mirrored change feeds to real-time intelligence. This eliminates the need to write custom Spark notebooks to poll for incremental updates and use Eventstreams SQL and no-code operators to process changes and build event-driven applications using Activator.
- Discover CDF-enabled mirrored databases in the Real-Time Hub and connect to an Eventstream in a few clicks — no code required.
- Stream row-level inserts, updates, and deletes with full fidelity, preserving the source table structure and change metadata.
- Apply Eventstream processing — SQL operators, filtering, and aggregation — and route outputs to Eventhouse, Activator, Lakehouse, and other destinations simultaneously.
- Works with all sources supported by Fabric Mirrored Databases, including Azure SQL, Cosmos DB, Oracle, PostgreSQL, Snowflake, and Open Mirroring partners.
Figure: Streaming mirrored database change events through Eventstreams for real-time processing and analytics.
Prerequisite: Delta Change Data Feed must be enabled on your mirrored database through Extended Capabilities.
For more information about this feature, refer to the Building real-time, event-driven applications on Mirrored Database Change Feeds with Fabric Eventstr... documentation.
Eventstream Observability with Workspace Monitoring (Preview)
Fabric Eventstream now integrates with Workspace Monitoring, giving you queryable observability into your streaming pipelines.
When Workspace Monitoring is enabled, Eventstream automatically emits node status, throughput metrics, and error data into your monitoring Eventhouse as KQL tables. This means you can track event ingress and egress across each Eventstream component, track system health, and analyze historical trends — all using KQL queries.
Three tables are now available:
- EventStreamNodeStatus: pipeline topology and node health.
- EventStreamMetrics: throughput, volume, and latency per node.
- EventStreamErrorMetrics: error counts and categories.
- A fourth table for diagnostic logs is coming soon to enable root-cause troubleshooting with actionable error messages.
With this integration, you can build real time dashboards over your streaming data, configure alerts on pipeline health, and run cross-item analysis across all Eventstreams in a workspace — all from a single Eventhouse.
Learn more: Overview of Microsoft Fabric eventstreams
Eventhouse remote MCP (Preview)
Eventhouse remote MCP enables AI agents to securely query and analyze real-time data in Eventhouse using natural language and KQL. With the hosted remote MCP server, there’s nothing to install — just point your agent to your Eventhouse endpoint.
Agents can discover schemas, generate KQL, sample data, and return insights across real-time and historical data. This unlocks seamless integration with platforms like GitHub Copilot, Copilot Studio, and Azure AI Foundry for real-time AI experiences.
Get started with Eventhouse MCP.
Eventhouse OneLake now supports schema evolution
Eventhouse OneLake Availability creates a logical copy of the data in Eventhouse by making it available in OneLake as delta tables. You can now add or delete columns without turning off OneLake availability.
These operations no longer require disabling availability or risk data loss, enabling seamless schema changes. This makes ongoing schema updates simpler, safer, and fully non-disruptive.
Configure and manage Activator rules directly in Eventstream (Preview)
Eventstream supports creating and managing rules directly in the experience.
Previously, setting up an alert required switching from Eventstream to Activator. While powerful, this meant moving between experiences to complete a single workflow. Now, alert creation is embedded directly into Eventstream.
Capabilities when building or editing your Eventstream:
- Select the stream you want to monitor.
- Choose Set Alert.
- Define your condition (thresholds, aggregations, patterns)..
- Configure the action.
- Create the rule.
This_image_shows_how_you_can_set_alert_and_manage_them_inside_EventstreamThis_image_shows_how_you_can_set_alert_and_manage_them_inside_Eventstream
Figure: Set alert inside Eventstream.
If you have Activator destination created on Eventstream, you can:
- Select Activator node
- Select Rule icon
- Create the rule
Once you have the rule(s) created on your Eventstream, you can manage them by editing, deleting or opening in Activator.
Figure: Manage rules inside Eventstream.
To learn more, refer to the Add a Fabric activator destination to an eventstream documentation.
Custom CA and mTLS support in Eventstream connectors (Preview)
Fabric Eventstream under Real-Time Intelligence provides various streaming connectors, enabling the integration of real-time data from popular sources into Fabric. When the Eventstream connector client establishes a connection with sources, it is required to implement TLS or mTLS encryption to fulfil the necessary security standards.
Currently, Eventstream connector service only leverages the system predefined CA certificates issued by trusted CA list to have the TLS support. When the customers’ source systems CA certificates are not in the trusted CA list, or the source systems require mTLS, Eventstream connector is not able to connect to the source systems.
The ‘Custom CA and mTLS Support’ feature addresses this gap by allowing customers to specify their custom CA and client certificates that are managed in their own Azure Key Vault when configuring their source in Eventstream. Once specified, Eventstream connector will fetch certificates from customers’ Azure Key Vault and use them to encrypt the connection. Now available in preview for Kafka-based Sources (Apache Kafka, AWS MSK, and Confluent Cloud for Apache Kafka) and Confluent Schema registry.
Figure: Eventstream connector custom CA and mTLS support flow.
To learn more about the configuration, refer to the Eventstream sources overview page and choose the corresponding source.
That’s a wrap for April!
We hope these updates help you move faster, simplify your day-to-day workflows, and get more out of Microsoft Fabric. There’s a lot to explore this month across the platform—whether you’re building, operating, or scaling data solutions—and we’re looking forward to seeing how you put these capabilities to work.