Fabric Weekly
Issue 5
4th August 2026
🎉 Welcome to issue #5 of Fabric Weekly!
This week, on the official side:
- The Fabric July 2026 Feature Summary rounds up everything from the month in one place - free DP-600/DP-700 certification vouchers via Data Days, FABCON and SQLCON in Barcelona this September, plus a batch of new deployment, governance, Spark, Eventstream, and Real-Time Intelligence capabilities.
- OneLake security improvements for SQL analytics endpoints - signed-in identities now pass straight through to OneLake, so RLS/CLS/OLS defined once applies everywhere, with better nested-group and service-principal support.
- And if you want to see who's been doing standout work across the ecosystem, check out this month's Fabric Influencers Spotlight, highlighting MVPs and Super Users across all corners of Fabric.
From the community:
- Nikola Ilic makes the case to stop physically materializing your silver layer now that GPU acceleration has landed for Fabric Data Warehouse - the logical separation still matters, the physical tables maybe don't.
- And Jason Cockington digs into the hidden cost of VertiPaq scans - two DAX measures can return identical results while taking very different paths through the engine.
- Alongside the OneLake security update above: Aleksi Partanen has a good reminder that workspace admins bypass OneLake Security roles entirely - contributor, member, and admin all see everything regardless of role, so don't assume RLS/CLS/OLS is airtight without checking who sits above viewer level in your workspace.
And finally, Reza Rad's Fabric Insider series has continued with Hasan Abo-Shally on Fabric MCP Servers, CLI & AI Agents - demos include a full Medallion architecture provisioned from a single pipeline run, and a Power BI report authored end-to-end inside one agentic conversation. And Wee Hyong Tok joined for Episode 9 to talk through where ETL is headed in Fabric - Gen2 Dataflows, DBT, SCD Type 2 automation, and the Fabric Data Factory MCP Server.
🤖 Copilot, AI & Agents
- Insights into the DP-800 exam and the Microsoft Fabric connection The author shares personal experience passing Microsoft's new DP-800 (SQL AI Developer Associate) exam and explains its Fabric angle - though note DP-800 is a cross-platform cert spanning SQL Server, Azure SQL, and SQL databases in Microsoft Fabric, unlike DP-600/DP-700 which are Fabric-exclusive. Includes updated study resources.
- The Best Use of AI for the Least Money? It's Not Inside VS Code Ginger Grant reveals that using Claude Code Pro outside of VS Code is a more cost-effective way to integrate AI into Fabric development after GitHub Copilot's fee changes.
- Retirement of Fabric data agent integration in Copilot in Power BI The Fabric data agent integration will retire in Copilot for Power BI on August 26, 2026, due to the OpenAI Assistant API's upcoming retirement, though Fabric data agents will still be available via Foundry, Copilot Studio, and M365 Copilot.
- Bring Enterprise planning to the AI Era with Planning in Microsoft Fabric IQ Planning in Microsoft Fabric IQ unifies enterprise planning with analytics and AI enabling finance teams to create budgets forecasts and scenarios directly on governed data without exporting or reconciling across systems.
- Fabric data agents in Microsoft Foundry: Easier to connect, easier to trust Fabric data agents in Microsoft Foundry are now easier to discover, integrate, and monitor with Model Context Protocol integration, OneLake Catalog discovery, multi-agent support, and improved logging.
- Musings on Agentic Development in Fabric Brief reflections on lessons learned from building an agentic development project on Microsoft Fabric. (Source post is very short; treat this summary as provisional pending fuller content.)
📊 Reporting & Insights
- Agentic AI in Power BI and Fabric, Part 4: From Skills to an Agentic Report Theme Audit Workflow This article details how Power BI and Fabric leverage agentic AI to automate a report theme audit workflow, integrating skills and guardrails for enhanced data analysis.
- Analysing Power BI Data Using Copilot In Excel Power BI users can now leverage Copilot in Excel to perform analyses like ABC analysis directly on semantic model data, generating both raw transaction tables and formatted reports with minimal setup.
- See your business as it happens with Real-Time Dashboard in Microsoft Fabric Real-Time Dashboard in Microsoft Fabric provides instant, actionable insights from streaming data, enabling operators and analysts to respond within seconds rather than waiting for daily reports.
- Bring Real-Time Dashboards into your apps with Microsoft Fabric Embed (Preview) Microsoft Fabric Embed allows developers to integrate interactive real-time dashboards into web apps via the Fabric Embed SDK, preserving access controls and improving usability.
- The Hidden Cost of VertiPaq Scans The Hidden Cost of VertiPaq Scans shows that identical DAX measures can have significant performance differences affecting report efficiency unnoticed by users.
- Livestream: BI Pixie — Power BI Usage Analytics & AI Readiness (with Gil Raviv) BI Pixie offers detailed usage analytics with heatmaps and satisfaction signals, plus an engagement-weighted Readiness Score to assess AI model readiness in Power BI.
- Fabric MCP Servers, CLI & AI Agents — A Conversation with Hasan Abo-Shally | Fabric Insider Ep. 8 Hasan Abo-Shally explains that an MCP Server standardizes AI agent interactions with external systems, enabling seamless actions like file summarization and Azure storage backup, bridging conversational AI and real-world operations.
- DaxLib.SVG 1.0.1 DaxLib.SVG is a DAX UDF library for embedding performant SVG charts in Power BI tables, matrices, and cards; this post marks its v1.0.1 release, following the earlier 0.2.3 beta.
- Contributing to DaxLib A short guide on contributing new DAX user-defined function libraries to DaxLib, the community-driven repository (built on SQLBI's DAX Lib platform) for sharing UDF libraries used in Power BI and Fabric semantic models.
- SVG Rendering Cost in Power BI Testing SVG rendering cost in Power BI (prompted by a DAXLib session and Chris Webb's Performance Analyzer talk) found most element types (filters, gradients, animation) render similarly, but rendering many elements compounds cost — 10 tables of 10,000 circles each took 5 seconds to render, 7 seconds when combined into one table. Practical advice: optimise the DAX, reuse elements with defs/use, avoid animations and complex filters, and limit SVG visuals per page.
- DAX Performance Tuner MCP Server Using the DAX Performance Tuner MCP Server across three LLMs (Claude Sonnet 4.5, GPT-5.2, Gemini 3 Flash) to optimise DaxLib.SVG.Compound.Violin(), the GPT-5.2 variant came out fastest — 923ms vs the original 1907ms, roughly a 52% reduction — by dropping an unnecessary axis-index table scan. Takeaway: useful for spotting oversights, but still needs rigorous testing and human validation.
⚙️ Data Engineering
- Change Event Streaming for SQL database in Microsoft Fabric (Preview) Change Event Streaming for SQL database in Microsoft Fabric (Preview) enables near-real-time processing of inserts, updates, and deletes by streaming changes directly to Fabric Eventstream or Azure Event Hubs, allowing developers to build low-latency pipelines entirely within the Fabric environment.
- Secure data ingestion with COPY INTO and Workspace Identity (Generally Available) Fabric's COPY INTO feature combined with Workspace Identity lets organizations securely load external data into their warehouse without exposing sensitive source file permissions.
- Eventhouse for data engineers: A faster path from events to insights Eventhouse in Microsoft Fabric speeds up data engineering by handling high-volume event data in near real time, reducing complexity and latency.
- Data Integration, Migration & the Future of ETL in Microsoft Fabric — A Conversation with Wee Hyong Tok | Fabric Insider Ep. 9 Wee Hyong Tok discusses how Microsoft Fabric unifies data integration into four pillars-unifying assets to OneLake, orchestrating pipelines, transforming data with Gen 2 Dataflows and DBT, and distributing results-to reshape ETL workflows.
- The Most Important Layer in Medallion Architecture? Bronze. In medallion architecture, the bronze layer-where raw data lands-is crucial; getting it wrong can cascade issues to downstream silver and gold layers, according to both the narrator and guest Teemu.
- (Don’t) Stop materializing your silver layer! The article advocates abandoning the silver layer as physical tables in medallion architectures due to recent GPU acceleration in Fabric Data Warehouse, which diminishes materialization benefits while maintaining logical data layer separation.
- Fabric July 2026 Feature Summary The July 2026 Fabric update brings free DP-600/DP-700 certification vouchers through Data Days, plus FABCON and SQLCON in Barcelona (Sept 28-Oct 1) with 130+ sessions, alongside new deployment, governance, Spark, Eventstream, and Real-Time Intelligence capabilities.
- Control where your streaming data goes with Outbound Access Protection for Fabric Eventstream Outbound Access Protection for Fabric Eventstream enables workspace admins to control and block streaming data from going anywhere until they approve specific exceptions via data connection rules.
- Building Event-Driven Applications in Microsoft Fabric with Business Events and User Data Functions Microsoft Fabric’s new Business Events and User Data Functions enable developers to build scalable event-driven applications with ready-to-use code samples and proven architecture patterns.
- Never Half-Manage Your Infrastructure with Terraform Heini's rule emphasizes full management of infrastructure components to prevent breaking environments; choose a clear boundary and automate everything within it.
- Job concurrency and queue monitoring for Spark, now with a capacity-level view (Generally Available) The new Spark feature provides a capacity-level view to help identify whether job delays stem from workspace limits, capacity saturation, or contention among shared workspaces.
- Customer-managed key encryption now extends to Spark jobs in Microsoft Fabric Microsoft Fabric now allows customers to manage encryption keys for Spark jobs, ensuring full control over data protection during active processing alongside storage.
- Question Marks in Power Query In Power Query, the question mark operator helps prevent query failures from missing data and allows you to specify default values for null results, making your data transformations more robust.
- Faster Spark queries with the Native Execution Engine: New performance optimizations The Native Execution Engine introduces several automatic performance optimizations to Spark queries, enhancing speed without requiring code changes.
🏛️ Storage & Platform
- Microsoft Fabric Just Got a MASSIVE Update! The July update for Microsoft Fabric introduces new features including bulk publish mode, Spark 4.1 and Delta Lake 4.2, enhanced diagnostic tools, faster UDFs, materialized lake views with event-driven refresh, scalar UDFs for T-SQL migrations, and improved Real-Time Intelligence connectors.
- How to check table sizes in a Microsoft Fabric warehouse Use a PySpark notebook with notebookutils.fs.ls() to recursively sum file sizes in each table's OneLake folder for determining table sizes in a Microsoft Fabric warehouse.
- Schema-Enabled Lakehouses Are Actually Faster (Microsoft Explains) The schema-enabled lakehouse in Azure Synapse speeds up data visibility by eliminating the sync delay, letting tables appear instantly after landing thanks to an integrated catalog system.
- Free Fabric Mirroring Won't Stay Free (And Maybe It Shouldn't) The upcoming premium pricing for Microsoft Fabric’s free mirroring feature reflects its substantial infrastructure costs, according to Vesa.
- Microsoft recognized as a Leader in The Forrester Wave™: Data Lakehouses Microsoft Fabric was recognized as a Leader in Forrester’s Q3 2026 Wave for Data Lakehouses, highlighting its unified AI-enabled lakehouse platform integrated with the Microsoft ecosystem.
- Bring Azure Monitor and AWS Glue data to OneLake (Preview) Bring Azure Monitor telemetry and AWS Glue Iceberg tables directly into OneLake with minimal data movement, enabling unified analytics across cloud platforms.
- Cross-domain intelligence with Azure Monitor data in Microsoft Fabric (Preview) Azure Monitor data integrated into Microsoft Fabric enables organizations to quickly translate system health signals into actionable business insights, addressing the common challenge of understanding impact before it's too late.
- Microsoft's Honest Take: No Urgent Need to Migrate Your Lakehouse Ted from Microsoft explains that there’s no immediate urgency to migrate non-schema lakehouses, as both types remain supported and you can schedule the transition according to your own timeline.
🔐 Governance & Security
- Workspace Admins Bypass OneLake Security in Microsoft Fabric In Microsoft Fabric, workspace admins can bypass OneLake Security roles entirely, posing a significant security risk.
- Purview Protection Policies in Microsoft Fabric Purview Protection Policies in Microsoft Fabric let workspace admins manage settings while blocking access to sensitive data through combined protection policies and sensitivity labels.
- OneLake security improvements for SQL analytics endpoints Microsoft Fabric’s OneLake Security now passes signed-in identities to the SQL analytics endpoint, enabling centralized governance with role-based access control, nested group support, improved shortcut-backed tables, enhanced column-level security for groups, and broader service principal support for automated workflows.
- Dynamic Data Masking in Power BI Warns against the common DAX-measure pattern for masking sensitive fields in Power BI Import models — the underlying column is still queryable via Build permission, XMLA, or Analyze in Excel, so it isn't real security. Instead, split sensitive fields into a separate "satellite" table secured by RLS, with masked and unmasked duplicate rows per record (flagged by an isMasked column), so the security role — not a report visual — decides which version a user sees. Trade-off: a larger table and careful design to avoid grouping/totals issues.
- New OneLake security improvements for Microsoft Fabric New OneLake security updates in Microsoft Fabric let organizations define consistent, intuitive access controls at the data layer that automatically apply across all Fabric experiences, including preview Eventhouse and Graph Database engines.
💰 Management & Cost
- The Fabric API Rate Limits Are Annoyingly Low The Fabric API imposes surprisingly low request limits, causing frequent 429 errors for developers who script against its REST endpoints.
- AI Keeps Getting Terraform Wrong. Reading the Docs Is Faster. The video demonstrates that relying on AI for Terraform guidance can be less efficient than directly reading the latest documentation, according to Heini's experience.
- CICD for Fabric Data Teams A tour of Fabric CI/CD options from low-code to full API control: Deployment Pipelines alone lack branching, review, and fine-grained parameterization; adding Git Integration brings version control and PR review; dropping Deployment Pipelines for pure branch-based promotion adds environment parity and git-revert rollback; and for complex solutions, API tooling (Terraform, fabric-cicd, Fabric CLI) with a solution-per-repo structure gives scoped ownership, blast-radius containment, and agent-friendly repos. Recommendation: start with git integration for the cheap wins, add sophistication as complexity grows.
- Fabric-CICD Updates - Semantic Models, Parameters and Config Covers fabric-cicd v0.1.17 through v0.1.34: configuration-based deployment via config.yml, parameter file templates with an extend key for reuse across workspaces, dynamic parameterization (including cross-workspace item references), key_value_replace for precise JSON key replacements, and Semantic Model Binding — mapping semantic models to specific data connections, including on-premises gateways and multiple connections per model.
- Fabric-CICD - It's Official Fabric-cicd, the open-source CI/CD tool for deploying Fabric items, reaches v0.2.0 with official Microsoft support, a committed long-term roadmap, and deeper integration with the Fabric platform.