Fabric Weekly
Issue 7
18th August 2026
👋 Welcome to issue #7 of Fabric Weekly!
This week, on the official side:
- Pankaj Arora announces that Capacity Overview Events in Real-Time Hub are now generally available - monitor capacity conditions, trigger alerts and automate responses in near real-time.
- Freddie Santos on identity-based predicate filtering for SQL audit logs - cutting the noise in enterprise-scale auditing by focusing on the activity that actually matters.
- Misha Desai on data source routing in Microsoft Fabric data agents, now generally available - a two-stage orchestration process that matches natural language queries against source metadata, schema samples and example queries.
- Nevena Nikolic on 10 things to know about GPU-powered Query Acceleration in Fabric Data Warehouse (preview) - boosting analytical performance with GPUs, no SQL changes or hardware upgrades required, tackling CPU bottlenecks in large scans, joins, filters and aggregations.
- Anasheh Boisvert on monitoring Eventstreams with Workspace Monitoring (preview) - now with per-Eventstream control, capturing structured metrics and error data in dedicated KQL tables for real-time troubleshooting.
From the community:
- Wolfgang Strasser rounds up Fabric Runtime 2.0's move to general availability - if you haven't opted in yet, note the dates: Runtime 2.0 becomes the default in late September, and Runtime 1.3 moves to long-term support from 1 October.
- Aleksi Partanen explains Fabric's new Ontology feature as a "company brain" for AI agents - grounding agents in your organisation's real data model rather than raw tables.
- Pragati Jain goes beyond the basics on Row-Level Security and Purview Labeling - a proper governance layer for Fabric, not just ticking the RLS box.
- Kevin Chant lays out important Microsoft Fabric CI/CD decisions - why the choices you make early on simplify setup and save you complexity down the line.
Here are four upcoming events across the UK (two of them reminders from last time, in case you missed them):
- Wednesday 2 September, Birmingham - #PBIBRUM host Vlad Mihanta on Power BI Embedded, and Neil Rowlands on building a Power BI training programme.
- Tuesday 15 September, online - Devon & Cornwall Power BI on Semanticus, a VS Code workbench for testing and shipping semantic model changes, from Kane Snyder. They're also running a session on agentic BI development the following week (22 September) with Kurt Buhler.
- Reminder - Thursday 27 August, Salford - PBIMCR's Ghost in the Aether, a murder mystery built entirely on Microsoft Fabric, from David Mitchell and Adrian Critchley.
- Reminder - Tuesday 8 September, online - Devon & Cornwall Power BI on MCPs, Skills and Plugins, with Cristi Prifti.
🤖 Copilot, AI & Agents
- Data source routing in Microsoft Fabric data agents (Generally Available) Microsoft Fabric data agents use a two-stage orchestration process to route queries by matching natural language input against source metadata, schema samples, and example queries.
- Reusable AI Skills for Writing to Microsoft Fabric Lakehouse Tables from Python Notebooks and to reduce Capacity Units consumption The post introduces reusable AI skills for efficiently writing data to Azure Fabric Lakehouse tables from Python notebooks, reducing Capacity Units usage without compromising performance.
- RANCH: Data Agent Responses API Notebook Conversion Helper RANCH automates migrating Fabric Data Agent notebooks from the deprecated Assistant API to the Responses API using regex-based code cell analysis and provides MCP endpoint guidance for external evaluation.
- How to Implement Retrieval in RAG Implementing effective retrieval in RAG is crucial for improving LLM response quality through targeted relevant information use.
- How to Implement Augmentation in RAG Augmentation in Retrieval-Augmented Generation structures prompts with system instructions, retrieved context, and the user's question to guide LLMs toward accurate, context-aware responses while managing token limits and document formatting.
- Microsoft Fabric Ontology Explained: A Company Brain for AI Agents The video explains how Microsoft Fabric’s new ontology item creates a “company brain” that lets AI agents understand complex business relationships, setting it apart from traditional semantic models.
- "Yeah, That's Oracle." The AI Fail That Made Her Switch to the MS Learn MCP The video shows how an AI model mistakenly provided Oracle SQL syntax for SQL Server 2025, prompting the creator to switch to the MS Learn MCP server for more accurate, learn-grounded responses.
- Build end-to-end Fabric solutions with SQL database skills (Preview) SQL database expertise now enables AI coding agents to seamlessly build, integrate, and maintain full Fabric workflows using familiar relational database operations.
- Sentiment analysis en temps réel avec Claude et Microsoft Fabric [French] This video demonstrates a real-time sentiment analysis pipeline using Claude’s API and Microsoft Fabric, optimizing cost and latency while integrating alerts via Teams.
📊 Reporting & Insights
- Optimising DAX: Data Materialisation Data materialisation in DAX occurs when the storage engine cannot preprocess data efficiently, leading to slow queries due to costly reconstruction, decompression, and poor memory utilization of entire tables by the formula engine.
- Optimising DAX: Practical Examples The article shows how small DAX tweaks, such as using bitmap filters instead of table materialization, greatly boost Power BI model performance while offering advice to prevent expensive duplicate scans and hidden costs from slicers.
- Capacity Overview Events in Real-Time Hub (Generally Available) Capacity Overview Events in Real-Time Hub’s general availability lets organizations monitor capacity, set up instant alerts, and automate responses within seconds.
- Slick Cars Visuals in Power BI Learn to create advanced AI-enhanced HTML card visuals in Power BI displaying last year's sales, growth percentages, dynamic colors, sparklines, and top products while verifying AI-generated DAX and HTML code for dashboard accuracy.
- On-premises data gateway July 2026 release The July 2026 on-premises data gateway release version 3000.326 improves manageability, visibility, and governance for connecting on-premises data sources to Power BI Fabric services.
- Creating DAX functions with AI to remove duplicated code The article demonstrates how AI can automatically refactor DAX code, eliminating duplication and promoting reusable user-defined functions for streamlined reporting.
- Displaying The Output Of Detail Rows Expressions In Power BI Reports Using The Paginated Report Visual This guide shows how to display Detail Rows Expression outputs in Power BI paginated reports for drill-through capabilities akin to Excel PivotTables.
⚙️ Data Engineering
- 10 Things to know about GPU-powered Query Acceleration in Fabric Data Warehouse (Preview) Query Acceleration in Fabric Data Warehouse boosts analytical performance using GPUs without changing SQL code or needing hardware upgrades, addressing CPU bottlenecks in large scans, joins, filters, and aggregations.
- Publishing Business Events from Fabric workloads Microsoft Fabric developers should publish business events from the specific workload that first detects a condition, ensuring signals originate near the source of truth to maintain data integrity.
- Fabric Runtime 2.0 is General Available Fabric Runtime 2.0 introduces major upgrades including a vectorized CSV parser, Delta Lake’s Liquid clustering, new Spark SQL features like the Arbitrary State API v2 and VARIANT type, alongside upgraded component versions such as Apache Spark 4.1, Java 21, and Azure Linux 3.0.
- Introducing Fabric Lakehouse support in dbt job for Microsoft Fabric (Preview) Fabric Lakehouse support in dbt jobs streamlines analytics engineering by enabling direct execution of dbt transformations on Microsoft Fabric's lakehouse architecture.
- I Completely Vibe Coded This Fabric Notebook with Claude Code In a casual demo, the creator shows how Claude Code was used to generate both the Fabric Notebook’s functionality and its documentation, though thorough validation has not been performed.
- Fabric Runtime 2.0 (Generally Available) Fabric Runtime 2.0 introduces a high-performance, Apache Spark-based engine that streamlines data engineering and science workflows within the broader Microsoft Fabric suite.
- Schedule Parameters: What Took You So Long, Microsoft? Mark from Microsoft explains that schedule parameters in Azure Data Factory took years to develop because the solution needed to scale for customers running billions of scheduled activities monthly.
🏛️ Storage & Platform
- Fabric as a Backend – Ep.554 Fabric as a Backend (FAB) streamlines app development by integrating workspace identity, Git, and SQL databases, reducing manual wiring typically required when using Azure.
🔐 Governance & Security
- Monitor your Eventstreams with Workspace Monitoring — Now with per-Eventstream control (Preview) Eventstream observability in Microsoft Fabric now allows users to toggle monitoring on or off for individual event streams, automatically capturing structured metrics and error data in dedicated KQL tables for real-time troubleshooting.
- A Practical Guide to use OneLake Security for OneLake Object, Row, and Column Security This guide shows how to configure OneLake security in a Fabric Lakehouse by removing broad-access roles and applying object-, row-, and column-level access for user-specific data visibility.
- Beyond the Basics: Deep Dive into Microsoft Fabric Row-Level Security (RLS) & Purview Labeling This deep dive explains how to implement Row-Level Security (RLS) and enforce Microsoft Purview sensitivity labels in Fabric for granular data protection.
- SQL Audit Logs: More Signal, Less Noise with Predicate Filtering (Generally Available) Predicate filtering in SQL audit logs reduces noise by focusing on relevant user activities, enhancing signal quality for enterprise audits.
💰 Management & Cost
- Important Microsoft Fabric CI/CD decisions This article outlines crucial CI/CD decision factors for Microsoft Fabric projects, emphasizing that early choices simplify setup and reduce long-term complexity.
- Understanding Warehouse Consumption with Capacity Metrics and Query Insights This blog explains how to use Capacity Metrics and Query Insights together to attribute warehouse consumption data at the query level, aiding in identifying workload drivers for CU usage and informing optimization decisions.
- Our Final Verdict on Terraform + Microsoft Fabric The verdict is that while Terraform integrates well with certain aspects of Azure’s Fabric, it should be used selectively rather than for full-scale management due to its specialized strengths.
📅 Community & Events
- Birmingham Power BI Meetup (#PBIBRUM): Power BI Embedded & Building a Training Programme — Wednesday 2 September, Birmingham Birmingham Power BI Meetup (#PBIBRUM) September session: Power BI Embedded, and building a Power BI training programme.
- Devon & Cornwall Power BI User Group: Build, Test and Ship Semantic Models with Semanticus — Tuesday 15 September, online Devon & Cornwall Power BI User Group: build/test/ship semantic models with Semanticus, plus a second September session on agentic BI development.
- Fabric & Power BI Manchester (PBIMCR): Ghost in the Aether — a Murder Mystery Built on Microsoft Fabric — Thursday 27 August, Salford Reminder: PBIMCR's Fabric-powered murder mystery event and a Lakehouse scaling talk, Thursday 27 August in Salford.
- Devon & Cornwall Power BI User Group: MCPs, Skills and Plugins — Tuesday 8 September, online Reminder: Devon & Cornwall Power BI User Group's MCPs, Skills and Plugins session, Tuesday 8 September online.