Azure Weekly
Issue 566
7th June 2026
Not only is this issue a Microsoft BUILD 2026 special; a bumper edition featuring all the announcements, but it also sports a new look - something I've been working on for the last few years - not only to support the ever increasing volume of Azure content, but to also better support the multiple newsletters we publish: Power BI Weekly and soon Fabric Weekly.
It's a packed edition (with 366 posts), but here are the high-level "what's new" posts:
- Microsoft Build 2026: Building agentic apps with Microsoft Fabric and Microsoft Databases by Arun Ulag - Introducing Rayfin, an open-source SDK and CLI that enables developers to define and deploy enterprise‑grade application backends directly into Fabric, accelerating the transition from prototype to production for agentic AI apps.
- Azure Functions at Build 2026 Update by Thiago Almeida - Azure Functions at Build 2026 introduces a first‑class programming model for AI agents, allowing developers to define and deploy serverless agents with markdown instructions in a .agent.md file across any trigger type while leveraging managed MCP tool servers, sandboxed code execution, and extensive connector support.
- What's new in Azure Container Apps at Build'26 by Vyom Nagrani - Azure Container Apps at Build'26 introduces Azure Container Apps Sandboxes, a public preview that offers fast, secure, ephemeral compute environments with built-in suspend and resume capabilities for running untrusted code safely and efficiently.
- What's new in Azure Kubernetes Service at Microsoft Build 2026 by Cory Skimming - Azure Kubernetes Service at Microsoft Build 2026 introduces managed system node pools for automatic lifecycle management, standardized Azure Container Linux to reduce patching overhead, bare metal AKS for direct hardware access, fleet manager for unified multi‑cluster operations, and Anyscale on Azure for integrated Ray orchestration of distributed AI
- Announcing new security, maintenance and analytics features for PostgreSQL at Microsoft Build 2026 by Guy Bowerman - The announcement at Microsoft Build 2026 introduces a suite of new security, maintenance, and analytics features for Azure Database for PostgreSQL flexible server, including the V6 SKU with NVMe storage, pg_duckdb and pg_ivm extensions for advanced analytics, Defender security assessments, temporal_tables for change tracking
- What's new in Azure API Management at Microsoft Build 2026 by Beena More - Azure API Management at Build 2026 introduces AI gateway enhancements, an Azure API Center that centralizes discovery and governance of APIs, agents, MCP tools, and AI assets, plus support for JSON‑RPC Agent‑to‑Agent APIs with content safety controls to manage emerging agentic workloads.
- What’s new in Observability at Build 2026 by Priyanka Nanda - Azure Build 2026 introduces AI‑focused agent observability in Azure Monitor, integrating OpenTelemetry for portable signals, new Agent Observability features with fleet views and automated evaluations, and an expanded Copilot Observability agent that enhances natural language investigation across AKS, Application Insights, and Foundry AI
- Azure Monitor Copilot Observability Agent: What’s new at Build by Efrat Nauerman - The updated Azure Monitor Copilot Observability Agent at Build 2026 introduces expanded investigation scenarios, integration with Microsoft Foundry AI Agents for cross‑scenario signal correlation, and enhanced capabilities in Application Insights, AKS clusters, Activity Logs, and future support for multiple Application Insights resources and Azure Service Health events
- What’s new in Microsoft Sentinel: May 2026 by Sowmy Srinivasan - May's update introduces unified role‑based access controls with row‑level scoping for granular permissions across Sentinel and Defender, expands connector catalog to over 400 integrations via Codeless Connector Framework, and previews Agent 365 for AI agent telemetry in the data lake.
Finally, it's worth highlighting Mark Russinovich's always fascinating Inside Azure Innovation session.
🤖 AI
- Generally Available: Microsoft Discovery Microsoft Discovery offers a unified platform for R&D teams to design, manage, and scale agentic AI workflows seamlessly across diverse fields, enhancing collaboration and governance.
- Generally Available: Microsoft Foundry IQ Microsoft Foundry IQ's general availability introduces a managed knowledge layer that seamlessly integrates enterprise data sources like SharePoint and OneLake, enabling developers to build AI agents without customizing retrieval pipelines for every project.
- Generally Available: Private Connectivity for Azure AI Search and Foundry Knowledge Bases Azure AI Search and Foundry Knowledge Bases now offer private, end-to-end network connectivity via Shared Private Link or Virtual Network peering, enhancing security and reducing latency for data flows between services.
- Generally Available: Text Analytics for Health NextGen Playground in Azure AI Language Azure AI Language's new Text Analytics for Health NextGen Playground lets users analyze complex medical texts—like discharge summaries and research abstracts—with high accuracy, enabling better clinical insights and AI model training.
- Generally Available: Self-serve custom photo avatar creation in Microsoft Foundry With the General Availability of self‑serve custom photo avatar creation, enterprise users can now quickly generate branded avatars for voice assistants and multimedia projects directly within the Foundry NextGen fine‑tuning portal.
- Generally Available: Custom Voice portal experience in Microsoft Foundry With General Availability, Azure AI Speech lets approved Custom Voice users directly manage recordings, consent forms, and training within the intuitive Microsoft Foundry interface.
- Generally Available: Unified Text Translation API in Azure AI Translator The Unified Text Translation API in Azure AI Translator now offers developers a single endpoint for various translation models, including neural, large and small language models, and adaptive custom translations.
- Generally Available: Voice Live integration with Microsoft Foundry Agent Service The newly GA Voice Live integration with Microsoft Foundry simplifies real-time speech processing by eliminating the need for custom audio pipelines, enhancing developer productivity.
- Generally Available: GenAI prompt skill and chat completion in Azure AI Search knowledge sources Azure AI Search now offers a General Availability feature enabling indexers to directly invoke chat completion models via the GenAI prompt skill, streamlining enrichment processes and eliminating the need for custom code.
- Generally Available: Managed identity for Foundry billing calls from Azure AI Search Azure AI Search introduces managed identity support for billing calls to Foundry, eliminating the need for key-based authentication and enhancing security through system‑assigned or user‑assigned identities.
- Generally Available: Azure AI Translator SDKs for the latest text translation API Azure AI Translator now offers General Availability SDKs for its latest text translation API, supporting C#, Java, JavaScript, and Python developers with seamless integration.
- Generally Available: Region-agnostic reservations for Global PTU in Microsoft Foundry With generally available region-agnostic reservations, customers can pre-purchase Global PTU capacity that works across all supported regions, simplifying workload scaling in Microsoft Foundry.
- Generally Available: LLM Speech API in Azure AI Speech Azure AI Speech's LLM Speech API now offers general availability, providing advanced LLM-powered transcription and translation in over 25 languages with enhanced accuracy and access across more regions.
- Generally Available: Custom Avatar and Custom Video portal in Microsoft Foundry The General Availability of Custom Avatar and Custom Video portals in Microsoft Foundry lets approved users create personalized media assets directly within the platform using Azure AI Speech technology.
- Generally Available: Content Understanding extractionMode for documents The new extractionMode in Microsoft Foundry's Content Understanding tool lets users switch between layout-aware analysis for structured docs and rapid text-only processing, optimizing performance based on document type.
- Generally Available: Speech SDK 1.50 for Azure AI Speech Azure AI Speech's Speech SDK 1.50 GA introduces enhanced multilingual support, including new models for over 120 languages and improved accuracy in noisy environments.
- Generally Available: Photo Avatar standard and custom in Azure AI Speech Azure AI Speech now offers both standard and custom Photo Avatar capabilities, enabling users to generate talking head videos from photos or tailor the avatar's appearance using their own images.
- Generally Available: Adaptive custom translation in the Azure AI Foundry NextGen playground Adaptive Custom Translation in Azure AI Foundry NextGen lets businesses tailor translations with just a few domain-specific examples, reducing reliance on large training datasets.
- Generally Available: Document translation for image files (synchronous, single document) Azure AI Translator's new GA feature lets users translate entire image files—such as JPEGs and PNGs—in a single synchronous API request, delivering instant results.
- Generally Available: Image translation inside Office documents for batch document translation Azure AI Translator has introduced general availability for batch document translation that can now process and translate text embedded within images in Word .docx files, automating the detection, OCR, translation, and re-rendering of image regions.
- Generally Available: Image file translation for batch document translation in Azure AI Translator Azure AI Translator now offers general availability for translating batch document images in formats like JPEG, PNG, BMP, and WebP, automating OCR before translation.
- Generally Available: Improved PDF batch document translation in Azure AI Translator Azure AI Translator now offers general availability for enhanced PDF batch document translation, leveraging Document Intelligence to accurately process both digital and scanned PDFs with preserved structure.
- Public Preview: Voice sync for avatar support in Voice Live API in Microsoft Foundry In the public preview of Voice Live API within Microsoft Foundry, developers can now integrate custom-branded avatars with synchronized voice output, enhancing immersive communication in applications.
- Public Preview: Tool search in Microsoft Foundry toolboxes The public preview of tool search in Microsoft Foundry's toolboxes lets users quickly locate specific tools within large, multi‑team catalogs, improving efficiency as the number of definitions scales.
- Public Preview: Content Understanding chunking and image verbalization in Azure AI Search Azure AI Search's public preview introduces Content Understanding chunking and image verbalization, enabling indexers to automatically divide complex documents into coherent segments and describe images textually for enhanced search capabilities.
- Public Preview: MAI-Voice-2 in Microsoft Foundry Microsoft Foundry's public preview of MAI‑Voice‑2 introduces a first‑party voice model capable of generating natural speech in over 10 languages, including voice cloning from brief reference samples.
- Public Preview: User feedback logging in Microsoft Foundry The public preview of User feedback logging in Microsoft Foundry allows developers to collect and analyze structured user interactions like thumbs ups, ratings, and custom notes directly within AI agent workflows.
- Public Preview: Agent-to-agent (A2A) support for Prompt agents and Hosted agents in Foundry Foundry introduces agent‑to‑agent (A2A) support in public preview, enabling Prompt and Hosted agents to securely invoke one another via a managed endpoint for streamlined workflow orchestration.
- Public Preview: Conversational PII NextGen Playground in Microsoft Foundry The new ConvPII NextGen Playground in Microsoft Foundry lets developers test and configure PII detection tools directly within a conversational interface, streamlining compliance for complex multi-speaker dialogues.
- Public Preview: TextPII NextGen Playground updates in Microsoft Foundry The new TextPII NextGen Playground in Microsoft Foundry's public preview introduces a revamped API Configuration Panel and previews Ignite 2025's PII categorization tools, enabling developers to experiment with automated data redaction.
- Public Preview: Domain filter for specialized model discovery in Foundry model catalog The public preview introduces a domain filter in Microsoft Foundry's model catalog, enabling developers to quickly locate over 1,900 specialized models—ranging from robotics to biomedical sciences—for targeted applications.
- Public Preview: Rubric evaluator in Microsoft Foundry The public preview of Rubric evaluators in Microsoft Foundry lets developers set custom quality metrics for agents through prompts, enhancing both single‑turn and multi‑turn workflows.
- Public Preview: Evaluations with Intelligent Trace Sampling Microsoft Foundry's new intelligent trace sampling feature lets engineers run evaluations on a representative subset of production traces, reducing data volume by up to 90% while maintaining accuracy.
- Public Preview: Microsoft Purview sensitivity label auditing in Azure AI Search Azure AI Search's public preview introduces audit events for sensitivity label actions, enabling organizations to track how Purview labels influence search results and compliance workflows.
- Public Preview: Serverless indexers in Azure AI Search Azure AI Search's new serverless indexers eliminate the need to provision compute, automatically scaling execution based on workload demands to reduce costs during idle periods.
- Public Preview: Hosted Agents in Microsoft Foundry Agent Service The public preview of Hosted Agents in Microsoft Foundry allows developers to run custom agent code in managed per-session VMs, simplifying deployment across frameworks.
- Public Preview: Purview sensitivity labels in Azure AI Search knowledge sources Azure AI Search now offers public preview of Purview sensitivity labels, enabling secure classification of data as it moves from source systems to search indexes and supporting Foundry agents and copilots.
- Public Preview: Purview admin access auditing for sensitivity-labeled content in Azure AI Search Azure AI Search introduces a public preview of Purview admin access auditing, enabling real-time tracking of sensitive content modifications across search, knowledge retrieval, and Foundry IQ platforms.
- Public Preview: SharePoint in Microsoft 365 indexer supports ASPX pages and Lists The updated Azure AI Search SharePoint indexer can now process ASPX pages, Lists, and document libraries, streamlining content discovery across multiple SharePoint formats.
- Public Preview: Incremental SharePoint permissions sync for Azure AI Search and Foundry IQ The public preview of incremental SharePoint permissions sync for Azure AI Search and Foundry IQ keeps document-, site-, and library ACLs up to date automatically as SharePoint permission changes occur, streamlining compliance and reducing manual updates.
- Public Preview: CLI and SDK support for Deployment Templates Azure Machine Learning introduces CLI and SDK support for deployment templates in public preview, enabling platform teams to manage these templates programmatically within infrastructure‑as‑code workflows using az ml commands.
- Public Preview: Deployment Templates support packaging deployment configurations in registries for curated deployment Azure Machine Learning's new public preview lets deployment templates pull configurations directly from registries, streamlining curated deployments and supporting a wider range of model sources.
- Public Preview: Advanced filtering for DiskANN vector search in Azure HorizonDB Azure HorizonDB's public preview introduces advanced filtering for DiskANN vector search, enabling simultaneous use of vector similarity and metadata criteria like category, date range, price, or tenant ID while maintaining high performance.
- Public Preview: Benchmark evaluations for fine-tuned models in Microsoft Foundry Microsoft Foundry's public preview includes benchmark evaluations for fine-tuned models, helping users select the optimal AI system by comparing performance across reasoning, math, and domain-specific tasks.
- Public Preview: Neural HD voice update (HDv2.5) in Azure AI Speech Azure AI Speech introduces HDv2.5, boosting audio fidelity by 30%, supporting expressive voice tags for enhanced narration, and cutting word error rates by 15% compared to the previous DragonHDLatest model.
- Public Preview: Trace Replay and trace visualizations for Foundry agents Foundry's public preview introduces Trace Replay and enhanced visualizations, enabling developers to efficiently debug complex multi‑step workflows by stepping through captured traces.
- Public Preview: Image serving in Foundry IQ knowledge bases Foundry IQ's new public preview lets users retrieve document‑extracted images—such as figures, screenshots, and diagrams—at query time, enhancing knowledge base searches with visual context.
- Public Preview: Foundry Memory preview refresh The refreshed Foundry Memory public preview introduces a redesigned storage architecture, enhanced retrieval management features, and improved compatibility with the Foundry Agent Service, enabling agents to perform write, search, and update operations on user-specific data.
- Public Preview: Content Understanding NextGen in Microsoft Foundry In the public preview of Microsoft Foundry, Content Understanding NextGen integrates directly into the portal, eliminating the need to switch between tools and providing developers with built-in Read and Layout playgrounds for seamless OCR text extraction.
- Public Preview: MCP Server Knowledge Source in Microsoft Foundry IQ In the public preview of Microsoft Foundry IQ, teams can federate knowledge from external MCP servers—including Salesforce, Atlassian, Confluence, and ServiceNow—to enhance retrieval capabilities beyond their indexed data.
- Public Preview: MAI-Transcribe-1.5 in Microsoft Foundry model catalog MAI-Transcribe-1.5 in Microsoft Foundry offers enhanced speech‑to‑text accuracy, particularly reducing word error rates for less common languages compared to earlier models.
- Public Preview: MAI-Image-2.5 in Microsoft Foundry model catalog MAI-Image-2.5 in Microsoft Foundry's public preview offers enhanced photorealism, better prompt adherence, and faster generation speeds over the previous MAI-Image-2 model.
- The 8192-Token Cliff: Why Your .NET RAG Pipeline Throws Random 500s The article explains why .NET RAG pipelines sometimes throw random 500 errors due to an unexpected 8,192‑token limit on embedding models, which can be avoided with proper length guards and truncation.
- How to stop AI hallucinations in enterprise RAG systems (a complete guide) This guide explains why Retrieval-Augmented Generation (RAG) systems often fail in enterprise settings—primarily due to poor chunking, weak embeddings, outdated documents, low‑confidence search results, and inadequate grounding techniques—while outlining five architectural patterns that high‑performing teams use to mitigate these
- Microsoft Just Made the Patient AI Conversation Official. Microsoft’s Copilot Health, now in preview, handles over 50 million daily health queries across consumer products, marking a significant shift as the largest enterprise software company officially embraces patient‑AI conversations.
- Building a Custom Video Captions Generator CLI Using MAI-Code-1-Flash in Copilot CLI The custom SRTgen CLI leverages MAI‑Code‑1‑Flash to streamline video caption generation, integrating Whisper and FFmpeg while offering flags for generating SRT files or embedding captions into MP4s.
- From insight to action: how Adobe and Microsoft are helping marketers move faster with AI Adobe and Microsoft are collaborating to embed customer experience intelligence directly into Microsoft 365 Copilot, enabling marketers to transition seamlessly from insight generation to execution within familiar tools like Teams, Word, and PowerPoint.
- Harness-Driven Agents: Secure Podcast Pipeline in Hyperlight MicroVM Sandbox Harness‑Driven Agents securely runs podcast generation agents in Hyperlight MicroVM sandboxes, preventing dangerous actions like rm -rf while enabling autonomous code execution for a daily Mandarin FIFA World Cup 2026 podcast.
- Use Your Mac for AI Agents: Self-Host Gemma 4 12 B with Pulumi and Tailscale Self-hosted Gemma 4 12B runs efficiently on a modern Mac using llama.cpp, offering local inference without data leaving your network and zero token costs, all managed via Pulumi and Tailscale for secure access.
- Accessibility Tools in Windows & Microsoft 365 Every Nonprofit Should Know This article highlights essential Windows and Microsoft 365 accessibility tools—such as Speech‑to‑text, Visual Assistance (magnification and Immersive Reader), Hearing Assistance (real‑time captions in Teams), and Cognitive Support features—that nonprofits can leverage to ensure all users, including those with disabilities, have
- Azure Speech at Build 2026: Powering Voice Agents with Real-Time and Life-like Experiences Azure Speech at Build 2026 introduces Voice Live for Foundry Prompt Agents, enabling developers to create enterprise‑ready voice agents with minimal overhead, along with public preview options for hosted agents and advanced models like GPT-Realtime 1.5 and Azure‑Realtime, enhancing responsiveness, mult
- Foundry Labs @ Build 2026 Foundry Labs at Build 2026 introduces a refreshed hub for AI experiments, featuring searchable catalogs, developer stories, community calendars, and organized categories across biomedical sciences, chemistry, code engineering, creative media, geospatial science, and robotics, along with new MAI models in text, image,
- Evaluate before you ship: introducing the Voice Live Evaluation Harness The Voice Live Evaluation Harness provides an open‑source, deployable pipeline that automatically scores multi‑turn voice interactions with Azure Voice Live using the same evaluators in Microsoft Foundry, enabling systematic quality baselines, configuration comparisons, regression detection, and data‑driven optimization across Semantic VAD, Push
- MAI-Image-2.5 support in ElBruno.Text2Image and the t2i CLI 🚀 The ElBruno.Text2Image .NET library and t2i CLI now support the new MAI-Image-2.5 and MAI-Image-2.5‑Flash models, enabling image generation directly from a terminal or C# applications with simple prompt‑to‑image workflows.
- Deploying Foundry Hosted Agents via REST API Deploying Foundry Hosted Agents via REST API simplifies infrastructure management by letting the platform handle compute, identity, and cross‑cutting concerns, enabling developers to focus on code while ensuring seamless integration with Azure services.
- Your first model deployment on Foundry Local on Azure Local: from catalog to inference in 10 minutes Foundry Local on Azure Local enables deploying open‑source models directly on your own Azure Local hardware with an OpenAI-compatible API, offering real‑time inference in about ten minutes using kubectl, Python, and a sample script.
- Announcing Microsoft Discovery general availability and Microsoft Discovery app preview Microsoft Discovery now generally available, offering a platform that integrates institutional knowledge, specialized tools, and review processes essential for complex scientific and engineering workflows, complemented by a new preview app to help teams quickly adopt agentic AI capabilities.
- A Developer’s Guide to Managing Models, Cost and Quality in Microsoft Foundry Microsoft Foundry introduces Fireworks AI on its platform, offering production‑grade open model inference with enterprise SLAs, while expanding the model ecosystem to include a wide range of modalities—from Microsoft AI models to partner and open‑source options—streamlining selection, evaluation, deployment, and operations for complex
- Foundry IQ: Build smarter agents faster with unified knowledge and serverless retrieval Foundry IQ simplifies building agent fleets by unifying enterprise and external knowledge sources, offering serverless retrieval with scale‑to‑zero pricing and instant context access via the Foundry IQ Model Context Protocol.
- OMO Teams: Multi-agent project delivery with ARB gates OMO Teams introduces an Architecture Review Board (ARB) gate system that uses dedicated AI agent teams for each phase of project delivery, ensuring governance, audit trails, and structured decision-making through YAML-configured voters and a Python script for tallying votes.
- Agents That Build Agents: A SKILL-first Blueprint with MS Agent Framework & Foundry The article introduces a two‑layer architecture where the Coding Agent builds an AI agent using a domain‑specific SKILL to translate business requirements into professional code, while the Runtime Agent operates the deployed system on platforms like Azure and GitHub Copilot, exemplified through ZavaShop's e-commerce use case.
- What's New in Microsoft 365 Copilot – May 2026 | 53 Updates Including Federated MCP Connectors The May 2026 update introduces federated MCP connectors for platforms like Canva, HubSpot, and Notion, enhancing cross‑service capabilities within Microsoft 365 Copilot.
🔎 Analytics
- Generally Available: OneLake catalog integration for Azure AI Search knowledge sources Azure AI Search now offers general availability of OneLake catalog integration, allowing users to register OneLake items once and reuse them across various knowledge sources and AI agents while maintaining original permission settings.
- Public Preview: Fabric IQ Ontology Knowledge Source in Microsoft Foundry IQ In public preview, Microsoft Foundry IQ integrates a Fabric Ontology as a federated knowledge source, enabling agents to query curated semantic layers directly from Fabric with natural language understanding.
- Public Preview: OneLake Catalog integration for Foundry IQ Knowledge Sources Foundry IQ's new OneLake Catalog integration lets agents retrieve authoritative data directly from lakehouse resources within Fabric, streamlining response grounding with a single registration step.
- Public Preview: Azure Databricks Lakebase branching with GitHub Copilot agent mode Azure Databricks introduces a public preview allowing developers to instantly branch a Lakebase production database, pair it with GitHub Copilot's agent mode for AI-assisted debugging on genuine production datasets, streamlining AI app development.
- Public Preview: Azure Databricks workspace-wide Genie MCP for Microsoft Copilot Studio The preview introduces a workspace‑wide Genie Management Control Plane (MCP) in Azure Databricks that lets Microsoft Copilot Studio agents ask natural‑language questions across the entire workspace via one endpoint, automatically routing queries to relevant Genie spaces.
- Playing Zork I as a Microsoft Fabric App via Rayfin Rayfin enables pro-coders to deploy NodeJS applications, like the classic text adventure Zork I, directly as Microsoft Fabric Apps, expanding real-time IoT analytics with interactive, low-code solutions.
- The Prompt Engineering Fallacy: Why Data Quality Still Determines AI ROI The article argues that data quality, not prompt engineering, determines AI ROI, using historical examples like Enron to caution against neglecting data foundations in favor of linguistic tricks.
- Turning Nonprofit Data Into Action: A Technical Guide to Power BI Power BI transforms scattered nonprofit data into unified, real‑time insights with its desktop modeling tools, cloud collaboration platform, and optional gateway for secure on‑premises connections, enabling advanced dashboards like multi‑program outcome reports and donor analytics powered by DAX.
- [Book Giveaway] The Definitive Guide to Microsoft Fabric The giveaway offers three community members the chance to win a free copy of a new Packt book on Microsoft Fabric, highlighting its practical insights for cloud professionals.
- From query acceleration to cache cooldown: Everything new in Fabric Data Warehouse at Build 2026 Fabric Data Warehouse’s Build 2026 release introduces query acceleration features and intelligent cache cooldown mechanisms to meet the rapid, high-volume demands of AI agents and copilots while maintaining data freshness and reliability.
- Data Factory parametrization via external mapping/config tables Data Factory now supports parametrization through external mapping and configuration tables, allowing users to bypass the 256-parameter limit and streamline pipeline management for complex projects.
- Graph in Fabric (Generally Available) Graph in Fabric’s general availability lets organizations explicitly model complex data relationships, enhancing AI systems' ability to understand connections and perform deterministic reasoning across large datasets.
- Dataflow Gen2 and Power Query innovations at Microsoft Build: Low-code data transformation with standout scale, performance, and reuse Dataflow Gen2 and Power Query now support low‑code data transformations that achieve standout scalability, superior performance, and enhanced reusability across Microsoft Fabric and Power BI.
- Benchmarking Dataflow Gen2: Faster data transformation at lower cost Benchmark results show Dataflow Gen2 in Microsoft Fabric achieves up to 10x faster data transformations at 30% lower costs than earlier versions, thanks to enhanced CI/CD pipelines integrated with Lakehouse and Power BI Direct Lake.
- Migration Assistant Enhancements for Fabric Data Warehouse The Migration Assistant for Fabric Data Warehouse now includes automated schema optimization and real-time performance analytics, making complex enterprise migrations smoother and more predictable.
- Excel Hell: 7 Signs You Need Data Automation Excel Hell’s seven signs—excessive manual tasks, version conflicts, high error rates, scalability issues, fragmented integration, inefficient reporting, and manual management of large datasets—indicate your business is ready for data automation to boost efficiency and accuracy.
- Advancing autonomous action with operations agent (Generally Available) Microsoft Fabric's operations agent autonomously monitors data, diagnoses issues, and performs necessary actions based on predefined goals, streamlining workflows beyond traditional dashboards and alerts.
- Orchestration in Fabric Data Factory: Build 2026 recap Fabric Data Factory's Build 2026 release introduces Approval activities for human decision-making, conditional retries for smarter resilience, and a modern canvas to simplify complex pipelines, enhancing both process orchestration and the people involved.
- Microsoft Purview data governance best practices Implementing notification processes for new data sources and integrating Purview into daily workflows are crucial best practices to ensure metadata trustworthiness and effective governance in addition to automating catalog updates.
- Driving AI-Powered Healthcare: Advanced Analytics, AI, and Real-World Impact Workshop The workshop explores how Microsoft Fabric unifies analytics to drive AI-powered healthcare solutions, emphasizing real-world impact through governance, security, and agent-based data activation.
- How to Dry Up Data Puddles Using the Data Fluency Framework The Data Fluency Framework offers a strategic approach to transform isolated "data puddles" into a unified information foundation, reducing operational costs and enabling AI models to scale by eliminating silos and promoting cross‑departmental data literacy.
- Consuming Azure blobs in a Fabric Notebook in realtime via an Eventstream This article demonstrates how to use an Eventstream and a Fabric Notebook to process Azure blob storage changes in real time, enhancing IoT solutions with instant data responsiveness.
- Fabric June 2026 Feature Summary Fabric June 2026 introduces enhanced core data engineering tools and warehouse features, alongside streamlined OneLake storage management and improved data agent observability.
- A new analytics frontier: GPU-accelerated Fabric Data Warehouse (Early Access Preview) The preview introduces a GPU-accelerated Fabric Data Warehouse designed to handle high concurrency and AI-driven workloads by reducing query latency and improving overall performance in dynamic analytics environments.
- From research to product: Microsoft Fabric wins best industry paper at SIGMOD 2026 for GPU‑accelerated data warehousing The winning SIGMOD 2026 paper details how years of research on running SQL on AI compute led to the development of a GPU‑accelerated data warehousing product in Azure’s Fabric platform.
- Analyze Your Business Events in Eventhouse and Real-Time Dashboards (Preview) Eventhouse lets you store and instantly analyze business events using KQL, while its preview version also supports building real-time dashboards for immediate insight.
- Fabric Skills for GitHub Copilot, Claude, and CLI: built by Microsoft, open for contribution Microsoft’s Fabric Skills open-source toolkit equips GitHub Copilot, Claude, Cursor, and Windsurf with accurate Fabric API usage, authentication, and workflow guidance for seamless integration within minutes.
- Building real-time, event-driven applications on Mirrored Database Change Feeds with Fabric Eventstreams (Preview) Fabric Eventstreams enables real-time, event-driven applications by streaming Mirrored Database Change Feeds directly into Fabric’s OneLake Delta tables, reducing latency to under 200 milliseconds for near‑live analytics and AI insights.
- Complex Data Processing with dbt Python Models: The Fabric Modern Data Platform This article demonstrates how to leverage dbt Python models in conjunction with Azure Databricks to perform complex data transformations, such as making REST API calls for data sanitization, expanding beyond what’s possible with standard SQL alone.
- Fabric IQ: The shared context layer for AI agents and real-time applications Fabric IQ provides a shared context layer that enables seamless real-time collaboration between human teams and AI agents by maintaining an up-to-date, unified view of business states.
- Microsoft Planetary Computer Pro is Generally Available Microsoft’s Planetary Computer Pro GA introduces enterprise‑grade reliability, supports modern geospatial formats like Zarr and NetCDF, and integrates with Fabric and Foundry to streamline geospatial AI development.
- Preview of multiparty analytics with Azure Confidential Clean Rooms Azure Confidential Clean Rooms introduces multiparty analytics using confidential compute-enabled Apache Spark SQL in a Trusted Execution Environment, enabling secure collaboration on privacy-sensitive datasets like PII and PHI across industries while protecting intellectual property and regulatory compliance.
- Build 2026: From data to intelligence—Faster with Fabric Data Factory Fabric Data Factory at Build 2026 streamlines data integration for AI applications, ensuring developers can confidently source trustworthy data and accelerate the delivery of intelligent solutions.
- Building the agentic analytics stack: Fabric Analytics at Build 2026 Fabric Analytics at Build 2026 introduces an end-to-end agentic analytics stack designed to embed AI directly into everyday user workflows, ensuring measurable business value is realized within familiar data tools.
- Deciphering Data Architectures Is Now Available in Multiple Languages Deciphering Data Architectures is now available in multiple printed languages including English, Portuguese, Polish, German, Russian, and Japanese, expanding its global accessibility for data professionals.
- Never ship a broken semantic model again: how to build automated tests in Power BI with user-defined functions User-defined functions (UDFs) in Power BI enable automated testing of semantic models with user-defined functions and PQL.Assert, reducing manual review time and ensuring consistent test standards across teams.
- Error Deploying GraphQL in Fabric: dm_exec_describe_first_result_set The article explains how a Fabric deployment error with dm_exec_describe_first_result_set was resolved by ensuring custom stored procedures used in a GraphQL object returned identical result set metadata across all code branches.
- Fabric Dataflows Gen2 And Concurrent Evaluation, Part 2: Partitioned Compute Fabric Dataflows Gen2’s Partitioned Compute feature shows how adjusting its Concurrency setting can dramatically speed up or slow down a single query with row‑level delays, turning a 1½‑minute refresh into over ten minutes when concurrency is reduced to one.
- Fabric Dataflows Gen2 And Concurrent Evaluation, Part 3: Is More Parallelism Always Good? While partitioned compute in Fabric Dataflows Gen2 boosts performance with higher concurrency, exceeding 16 parallel tasks fails due to SharePoint throttling, highlighting the need to balance parallelism for optimal results.
🖥️ Compute
- General Availability: Backup your ledger files for audit with Azure confidential ledger's new tool Azure now offers a tool that lets audit personas securely view and verify ledger files using its confidential ledger technology.
- Generally Available: OpenTelemetry Metrics, Visualizations, and Enhanced Monitoring in Azure Monitor for Azure VMs and Arc Servers Azure Monitor now offers general availability of OpenTelemetry metrics and visualizations, providing a unified monitoring experience with real-time insights for both Azure Virtual Machines and Arc Servers.
- Generally Available: Azure Functions now supports hosting MCP Apps Azure Functions now supports hosting MCP Apps, enabling developers to deliver rich, interactive experiences like visual data exploration and multi-step workflows directly within their serverless applications.
- Generally Available: Built-in Grafana dashboards for Azure Functions Azure Functions introduces ready‑to‑use Grafana dashboards that give developers instant visibility into app health, performance, and scaling metrics without needing to set up or manage a separate Grafana instance.
- Generally Available: Azure Functions Support for Node.js 24 Azure Functions now officially supports Node.js 24, enabling developers to build and deploy applications locally with the latest runtime while running on both Linux and Windows Azure Functions plans, including the flexible consumption option.
- Public preview: Azure Linux 4.0 for Azure Virtual Machines and VM Scale Sets Azure Linux 4.0 introduces a modernized stack with Linux kernel 6.18 LTS and dnf5 package manager, optimized for enhanced performance and security in Azure Virtual Machines and VM Scale Sets.
- Public Preview: Automatic OS Image Upgrades for VMSS Flex The public preview of automatic OS image upgrades for Azure VMSS Flex reduces manual workload by seamlessly updating entire fleets with consistent images.
- Public Preview: Go language support on Azure Functions Azure Functions introduces Go language support in public preview, enabling developers to build event‑driven serverless applications with native Go tools and libraries.
- Public Preview: Copilot-assisted coding & agent skills to build, deploy, and run Azure Functions faster in VS Code, CLI, and azd Azure's upcoming Copilot preview introduces agent skills that standardize AI assistance for coding, deploying, and running Azure Functions across VS Code, CLI, and azd, ensuring consistent high-quality support regardless of underlying model or data freshness.
- Public Preview: Serverless agents runtime for Azure Functions Azure Functions introduces a public preview of a serverless agents runtime that lets developers build and deploy AI agents directly within functions using a markdown‑first approach, enabling natural language event‑driven agent creation.
- Public Preview: Toolbox connectors and triggers in Microsoft Foundry The public preview of Toolbox connectors and time-based triggers in Microsoft Foundry enables event-driven AI agent development without manual invocation, streamlining automation workflows for enterprises.
- Public Preview: Snapshot backup for SQL Server in Azure VMs Azure Backup's new Snapshot feature for SQL Server on Azure VMs combines disk snapshots and transaction logs to provide nearly instant, low‑impact full backups even for large databases.
- Private Preview: Storage optimized Lasv5 and Laosv5 Azure VM series Azure's upcoming Lasv5 and Laosv5 VM series are optimized for storage, offering up to 40% higher disk I/O performance and lower latency for workloads demanding massive data handling.
- Private Preview: multiparty analytics with Azure Confidential Clean Rooms Azure Confidential Clean Rooms offers a private preview enabling multiple parties to perform secure, privacy-preserving analytics on sensitive data using Apache Spark within a fully managed environment.
- Preview: Azure Cobalt 200 Arm-based Dpsv7, Dplsv7, Epsv7, Mpsv4, and Lpsv5-series Virtual Machines Azure's new Arm-based Cobalt 200 series VMs—Dpsv7, Dplsv7, Epsv7, Mpsv4, and Lpsv5—provide up to 50% better performance and lower costs compared with the older Cobalt 100 lineup.
- Retirement: Azure VMware Solution AV36 node End of Support on September 30, 2027 Broadcom announces that the Azure VMware Solution AV36 node will reach end of support on September 30, 2027, aligning with VMware’s roadmap and becoming incompatible with VMware Cloud Foundation version 9.
- Query Public IP Addresses of Azure VMs Across All Subscriptions with Azure Resource Graph Azure Resource Graph enables you to efficiently retrieve comprehensive public IP details—including VM names, IP SKUs, allocation methods, and versions—for Azure virtual machines across all your subscriptions in a single query.
- Announcing Go support in Azure Functions (Preview) Azure Functions now offers first‑class Go support in public preview, enabling developers to build serverless applications with idiomatic Go, full trigger coverage across HTTP, Timer, Service Bus, Event Hubs, Cosmos DB, and Blob Storage, along with native Go build pipelines, integrated observability via Application Insights
- Introducing the Azure Functions serverless agents runtime (preview) The new Azure Functions serverless agents runtime simplifies building AI agents with a markdown‑first model, offering event triggers like timers and Teams messages while handling infrastructure details such as identity management and observability out of the box.
- Azure Functions MCP Extension: What's New at Build 2026 The Azure Functions MCP Extension now supports full MCP primitive sets—including tools, resources, and prompts—in multiple languages, enabling developers to build comprehensive remote MCP servers with built-in authentication and interactive UI capabilities via MCP Apps.
- Azure Functions at Build 2026 Update Azure Functions at Build 2026 introduces a first‑class programming model for AI agents, allowing developers to define and deploy serverless agents with markdown instructions in a *.agent.md* file across any trigger type while leveraging managed MCP tool servers, sandboxed code execution, and extensive connector support.
- Announcing Preview of new Azure Lasv5 and Laosv5 VMs based on the AMD EPYC™ ‘Turin’ processor The new Azure Lasv5 and Laosv5 VMs, powered by AMD EPYC “Turin” 5th Gen processors, offer up to 30.7TB of local storage for Lasv5 and 138TB for Laosv5, along with enhanced network bandwidth up to
- New Azure Cobalt 200 VMs deliver 50% performance improvement, fully optimized for modern agentic AI workloads Azure Cobalt 200 Arm-based VMs deliver up to 50% better generational performance, optimized specifically for modern agentic AI workloads with enhanced security, networking, and storage capabilities.
- REMINDER: Windows Server 2016 End of Life January 2027 – Plan Your AD CS Migration Now Plan ahead for migrating Active Directory Certificate Services from Windows Server 2016 before its January 2027 end-of-life to avoid compatibility issues and maintain security.
🚢 Containers
- Generally Available: Azure Red Hat OpenShift in Belgium Central The launch of Azure Red Hat OpenShift in Belgium Central expands European deployment options and aligns with Microsoft's growing presence in the region.
- Azure Container Linux (ACL) now generally available on Azure Kubernetes Service (AKS) Azure Container Linux (ACL) now generally available on AKS offers enhanced security and performance by using an immutable, container-optimized OS derived from Flatcar and tailored for Kubernetes workloads.
- Generally Available: Additional support for OpenTelemetry destinations (New Relic, Dynatrace, Elastic) Azure Container Apps now supports sending OpenTelemetry data to New Relic, Dynatrace, and Elastic, enabling developers to integrate their applications with these popular observability platforms more seamlessly.
- Generally Available: Monitor HTTP traffic in Azure Container Apps Azure Container Apps now offers a General Availability feature that captures detailed HTTP access logs at high volume, enabling better monitoring of application traffic through a new Azure Monitor diagnostic setting category.
- Generally Available: Override Scale Rules in Azure Functions on Azure Container Apps With the new allowScalingRuleOverride property, Azure Functions on Container Apps lets developers manually control scaling rules for better performance tuning and resource optimization.
- Generally Available: Confidential Compute support on Azure Container Apps Azure Container Apps now offers General Availability of Confidential Compute, enabling customers to run highly regulated workloads with hardware-enforced memory encryption and isolated processing environments.
- Generally available: Managed system node pools in AKS Automatic Managed system node pools in Azure Kubernetes Service (AKS) Automatic automate provisioning, scaling, patching, and high availability, reducing operational overhead for cluster administrators.
- Generally Available: Azure Kubernetes Fleet Manager for Arc-enabled clusters Azure Kubernetes Fleet Manager's general availability unifies management of hybrid and multi-cloud Kubernetes clusters into a single console with consistent tooling and operational models.
- Public Preview: Azure Container Apps Sandboxes Azure Container Apps introduces sandboxes that securely isolate untrusted code, simplify state management, and automatically scale to handle sudden traffic spikes for agentic apps and multi-tenant services.
- Public Preview: Anyscale on Azure Anyscale on Azure offers cloud-native managed Ray support, enabling ML engineers to run high-performance Python workloads at massive scale with full control and speed.
- Public Preview: Microsoft Defender for Cloud support for Azure Container Apps (Serverless Containers Posture) With the public preview, customers can monitor and enforce security policies for Azure Container Apps within Defender for Cloud's Serverless Containers Posture, streamlining workload protection across serverless containers.
- Govern AI Agents Using Agent Governance Toolkit and Azure Container App Sandboxes The Agent Governance Toolkit now integrates Azure Container App sandboxes to securely isolate and enforce policy on AI agent-generated code, preventing unauthorized access to secrets while providing hardware-isolated execution environments with sub-second startup times.
- Regional Endpoints for Azure Container Registry Geo-Replication — Now in Public Preview Azure Container Registry now offers public preview of regional endpoints for geo-replicated registries, allowing per-region login URLs with no feature flag or CLI extension needed, available on Premium SKUs across all Azure regions.
- Pod CIDR Expansion Generally Available and IP Address Planning on Azure CNI Overlay Pod CIDR Expansion in Azure CNI Overlay lets clusters dynamically adjust their address space up to 1,024 nodes without downtime, simplifying long-term IP planning and reducing the risk of overlapping CIDRs across clusters.
- AKS: Mount Versioned Content as OCI Image Volumes Kubernetes 1.36’s stable image volumes enable mounting any OCI artifact directly into pods as read‑only storage, eliminating the need for init containers or CSI drivers when deploying versioned project content via an AKS-hosted Go HTTP API.
- Build an EKS Environment Factory with Pulumi and vCluster The EKS Environment Factory with Pulumi and vCluster enables platform teams to provision isolated, ephemeral Kubernetes environments on demand, reducing provisioning time by 89% as demonstrated in Deloitte’s case study, while keeping costs low through shared infrastructure.
- PREVIEW: “Azure Linux 4.0” – Purpose-Built for Azure Azure Linux 4.0, a Fedora‑derived distribution optimized for Azure, offers a single, consistently patched, and lightweight OS that reduces operational overhead across VMs, Kubernetes nodes, and container images while integrating seamlessly with Azure Defender, Monitor, Migrate, and other services.
- AI-200 – Azure Container Registry – ACR Automatically Triggered Tasks Azure Container Registry’s Automatically Triggered Tasks enable continuous builds that start automatically with every code push to an Azure DevOps Git repository, eliminating manual CLI steps and ensuring images are instantly updated without human intervention.
- AI-200 – Azure Container Registry – ACR Quick Tasks Build Container Images Without Docker Desktop Azure's ACR Quick Tasks enables developers to build and push container images directly in the cloud without needing Docker Desktop or a local container runtime, streamlining CI workflows and eliminating version mismatch issues.
- Token economics–driven architecture: hybrid models, AI Runway, AKS Kata MicroVM, MCP This architecture uses AI Runway to tier models, place them on cost‑effective AKS nodes within Kata MicroVMs, and share a single small‑language model endpoint via the Model Context Protocol, dramatically reducing token costs for enterprise agentic workloads while preserving security.
- AI 200 – ACR Tasks: Build and Push Container Images Without Docker Desktop ACR Tasks eliminates the need for local Docker Desktop by automating container image builds, tests, and pushes directly in Azure, ensuring consistent CI/CD pipelines across developers without version conflicts or credential management issues.
- Inside ACR Artifact Cache: Pull-Through Caching at Scale Inside Azure Container Registry's Artifact Cache, pull-through caching enables platforms to map downstream repository paths to diverse upstream sources, ensuring consistent performance and security while offloading complex credential management and rate-limit handling from production workloads.
- Anyscale on Azure: Powering Enterprise AI at Massive Scale on Azure Kubernetes Service Anyscale on Azure offers a unified Ray runtime for end‑to‑end AI workloads, enabling seamless data preparation, training, fine‑tuning, inference, and agentic execution within a single Python environment on Azure Kubernetes Service, reducing fragmentation costs and improving GPU utilization to over 70%.
- Introducing Azure Container Linux (ACL) Azure Container Linux (ACL) combines Flatcar’s immutable security model with Azure‑specific binaries for streamlined integration in AKS, offering enterprise‑grade stability, reduced operational overhead, and enhanced supply‑chain trust.
- Announcing Azure Linux 4.0: Purpose-Built for Azure, Now in Public Preview Azure Linux 4.0 offers a unified, optimized Linux distribution for all Azure compute services, reducing operational complexity while maintaining high performance and security with seamless integration into Defender for Cloud, Azure Monitor, and other Azure tools.
- Introducing Azure Container Apps Sandboxes: Secure Infrastructure for Agentic Workloads Azure Container Apps Sandboxes provide secure, instant‑startup compute environments ideal for platform developers to build multi‑tenant solutions and AI agents to rapidly configure self‑contained workspaces with enterprise‑grade isolation.
- What's new in Azure Container Apps at Build'26 Azure Container Apps at Build'26 introduces Azure Container Apps Sandboxes, a public preview that offers fast, secure, ephemeral compute environments with built-in suspend and resume capabilities for running untrusted code safely and efficiently.
- Introducing On-demand Sandboxes for Azure Durable Task Scheduler (Private Preview) On-demand Sandboxes for Azure Durable Task Scheduler lets you run individual workflow steps in isolated, managed compute environments without managing infrastructure or orchestrator changes, supporting .NET and Python SDKs with per-activity isolation and scale‑to‑zero capabilities.
- Announcing Anyscale on Azure public preview: Powered by Ray on AKS Anyscale on Azure public preview offers a managed Ray platform running on AKS, simplifying distributed AI workloads by aggregating fragmented GPU capacity, unifying data across the AI lifecycle, and providing enterprise‑ready operations through familiar Azure tools.
- What's new in Azure Kubernetes Service at Microsoft Build 2026 Azure Kubernetes Service at Microsoft Build 2026 introduces managed system node pools for automatic lifecycle management, standardized Azure Container Linux to reduce patching overhead, bare metal AKS for direct hardware access, fleet manager for unified multi‑cluster operations, and Anyscale on Azure for integrated Ray orchestration of distributed AI
- Embed intelligence into physical systems with smaller form factor infrastructure (preview) Microsoft is previewing capabilities that enable embedding AI intelligence directly into physical systems through smaller form factor infrastructure, local agent execution with Foundry Local, and seamless integration with Azure Kubernetes Service and IoT Operations, aiming to simplify autonomous robotics development.
- The Best MCR Base Image for Your Language? SBI Scans Nightly So You Don’t Have To SBI automatically nightly scans MCR base images for vulnerabilities, ranking them by language and OS to help teams select the most secure and lightweight container foundation.
- Deploying the AKS Argo CD Extension with App Routing Ingress and Entra ID SSO The article provides a detailed walkthrough of deploying the AKS Argo CD Extension, integrating App Routing Ingress, securing TLS via Azure Key Vault, and enabling seamless Single Sign-On (SSO) using Entra ID along with advanced role‑based access control for workloads.
🗄️ Databases
- Generally Available: Azure Database for PostgreSQL - Flexible Server: DuckDB extension Azure Database for PostgreSQL now offers a DuckDB extension, enabling users to perform advanced analytics directly within their database environment with seamless integration and high performance.
- Generally Available: Agent kit for Azure Cosmos DB The Azure Cosmos DB Agent Kit provides developers with built-in best practices to optimize performance and simplify the creation of high‑throughput database applications in their AI coding workflows.
- Generally Available: Azure DocumentDB instant free tier clusters Azure DocumentDB now offers a generally available instant free tier for clusters, enabling developers to provision and start using the service without delay.
- Generally Available: Azure Database for PostgreSQL flexible server pg_ivm extension The newly generally available pg_ivm extension for Azure Database for PostgreSQL flexible server enables seamless interoperability between PostgreSQL and Snowflake, allowing users to share data across both platforms without code changes.
- Generally Available: Per-partition automatic failover for Azure Cosmos DB Azure Cosmos DB's new per‑partition automatic failover feature enables rapid recovery of individual database partitions without affecting the whole account, enhancing resilience for high‑availability workloads.
- Generally Available: Service-managed failover in Azure DocumentDB Azure DocumentDB's new service‑managed failover feature automatically switches multi‑region clusters to a healthy region within seconds during outages, ensuring continuous availability without manual intervention.
- Generally Available: Azure Database for MySQL now available in new regions and expanded availability zones Azure Database for MySQL's general availability expansion includes new regions like Denmark East and Austria East, along with expanded availability zones for improved reliability and lower latency.
- Generally Available: Azure Cosmos DB global secondary indexes Azure Cosmos DB's new globally available global secondary indexes let developers boost query speed with minimal code changes, as the indexes auto‑maintain synchronization across regions.
- Generally Available: Azure Cosmos DB all versions and deletes change feed mode Azure Cosmos DB's new all versions and deletes change feed mode provides real-time tracking of every container modification, enhancing data integrity and audit capabilities.
- Generally Available: MCP Toolkit for Azure Cosmos DB The MCP Toolkit for Azure Cosmos DB offers AI agents and copilots reliable access to operational data with enterprise-grade features now available in general availability.
- Generally Available: Change partition keys in Azure Cosmos DB for NoSQL API Azure Cosmos DB now offers a GA feature to change partition keys in NoSQL containers via the Azure portal, supporting both online and offline migration methods for seamless data transition.
- Generally Available: Azure Database for MySQL Flexible Server self-service quota management experience Azure Database for MySQL Flexible Server now offers a self‑service quota management experience via the Azure portal, allowing users to easily monitor usage and request adjustments without needing support tickets.
- Generally Available: Azure Cosmos DB Linux emulator (vNext) The Azure Cosmos DB Linux emulator (vNext) is now generally available, offering cross‑platform Docker support for local development and testing of NoSQL workloads with near‑identical behavior to the cloud service.
- Generally Available: Azure HorizonDB Agentic Advisor Solution Accelerator The Azure HorizonDB Agentic Advisor Solution Accelerator offers pre-built components for developing multi-agent financial applications in Python, streamlining integration with LangGraph and Mem0 while leveraging Azure HorizonDB's capabilities.
- Generally Available: Graceful failover with zero data loss guarantee Azure DocumentDB's new graceful failover capability promotes replica clusters without data loss, enhancing reliability over its previous force promote option.
- Public Preview: AI pipelines in Azure HorizonDB Azure HorizonDB's new AI pipelines feature lets developers define complex AI workflows—such as ingestion, chunking, embedding, extraction, generation, and ranking—in standard SQL, enabling fault‑tolerant execution directly within the database.
- Public Preview: Azure SQL as a knowledge source in Foundry IQ Azure SQL Database gains first-class status as a knowledge source in Foundry IQ's public preview, enabling seamless integration into Copilot, RAG, and agentic applications with built-in data authority.
- Public Preview: Advanced full-text search in Azure DocumentDB Azure DocumentDB's public preview introduces advanced full‑text search features, including fuzzy and proximity searches, expanded language support, and BM25 ranking, enabling unified search operations within a single database.
- Public Preview: Azure SQL updates for early-June Azure SQL now supports using AES 256 symmetric keys for customer‑managed transparent data encryption, enhancing security control for early-June 2026 users.
- Public Preview: Azure Backup for Cosmos DB Azure Backup for Cosmos DB's public preview introduces vaulted backups that ensure secure, isolated, and highly available protection for mission-critical data, meeting cyber‑resilience and regulatory compliance requirements.
- Public Preview: Change streams (multi-shard) in Azure DocumentDB Azure DocumentDB's new public preview of multi-shard change streams enables real-time, event‑driven applications with low latency polling for live data changes, ideal for instant analytics and automation.
- Public Preview: BM25 full-text search in Azure HorizonDB Azure HorizonDB's public preview introduces BM25 full-text search, delivering Elasticsearch‑ and Solr‑level relevance directly within the database for production-ready keyword queries.
- Public Preview: New Azure Cosmos DB cost estimator The new Azure Cosmos DB cost estimator in public preview lets users quickly model throughput, storage, regions, and capacity modes to plan deployments with confidence, reducing provisioning risks.
- Public Preview: Relational Database to Azure Cosmos DB NoSQL Migration Assistant The Azure Cosmos DB Migration Assistant in Visual Studio Code streamlines moving data from popular relational databases like Oracle, SQL Server, PostgreSQL, and MySQL with an intuitive, guided workflow.
- Public Preview: Agent Memory Toolkit for Azure Cosmos DB The Agent Memory Toolkit for Azure Cosmos DB lets developers create custom, code-first memory systems that boost agentic app scalability, durability, and cost efficiency during its public preview.
- Public Preview: Semantic reranker in Azure Cosmos DB The semantic reranker in Azure Cosmos DB uses natural language processing to automatically adjust query results, enhancing relevance without manual intervention.
- Public Preview: Microsoft Defender security assessments for Azure Database for PostgreSQL Microsoft Defender's public preview for Azure Database for PostgreSQL offers continuous security assessment, automatically identifying vulnerabilities and misconfigurations to help enhance database protection.
- Public Preview: Safe key rotation for Azure Cosmos DB Azure Cosmos DB's public preview introduces safe key rotation, which tracks the last usage of each account key to prevent disruptions when rotating credentials.
- Public Preview: Agentic Retrieval Toolkit for Azure Cosmos DB The Agentic Retrieval Toolkit for Azure Cosmos DB enables multi-hop, graph-based RAG applications with seamless integration, supporting complex query workflows out of the box.
- Public Preview: Azure Cosmos DB distributed transactions Azure Cosmos DB's public preview introduces distributed transactions enabling atomic cross-item and partition operations with automatic commit or rollback on failure, streamlining complex data workflows.
- Public Preview: MCP toolkit for Azure DocumentDB The MCP toolkit for Azure DocumentDB enables AI agents and LLMs to perform natural language queries on DocumentDB databases with low latency, facilitating seamless integration of generative AI into existing document storage solutions.
- Public Preview: Pre-upgrade validation checks for Azure Database for PostgreSQL Azure Database for PostgreSQL now offers public preview pre-upgrade validation checks to ensure compatibility and readiness before initiating a major version upgrade, reducing risk of downtime.
- Public Preview: Introducing AI Assistant in the Azure Cosmos DB extension for Visual Studio Code The preview introduces an AI Assistant within the Azure Cosmos DB VS Code extension, enabling users to formulate natural language queries that are automatically translated into optimized Cosmos DB commands, streamlining data interaction.
- Public Preview: Azure SQL Database provisioning in the MSSQL extension The MSSQL VS Code extension's new public preview lets developers provision fully managed Azure SQL Database instances in the free tier, enabling direct creation and connection to cloud databases straight from their code editor without any upfront cost.
- Public Preview: Azure Database PostgreSQL - Flexible Server cross-tenant customer-managed keys Azure Database for PostgreSQL - Flexible Server's public preview introduces cross‑tenant customer‑managed keys, enabling encryption with an Azure Key Vault key located in a different Entra tenant for enhanced security and compliance.
- Public Preview: OmniVec, an open-source embedding toolkit for Azure OmniVec lets developers automatically generate embeddings from diverse sources and stream them directly into Azure Cosmos DB and other Azure services, simplifying the pipeline for building AI applications.
- Public Preview: Azure HorizonDB Azure HorizonDB offers future‑proof autoscaling storage up to 128 TB, delivering performance three times faster than self‑managed PostgreSQL while simplifying database management.
- Public Preview: Integrated embeddings for Azure Cosmos DB for NoSQL Azure Cosmos DB for NoSQL's new public preview feature automatically generates and maintains embeddings, simplifying AI application development by eliminating the need for a separate embedding pipeline.
- Base64 Encoding and Decoding in SQL Server 2025 and Azure SQL Database SQL Server 2025 and Azure SQL Database now include native Base64 encoding and decoding functions, simplifying tasks like embedding binary data in JSON or creating URL‑safe tokens while avoiding the need for external libraries.
- A Multi-Region Microsoft Foundry Pattern for Enterprise Private Networking The multi-region Microsoft Foundry pattern enables enterprises to deploy Foundry where platform capacity exists while securely connecting to existing private resources, preserving security and operational standards across regions.
- Azure HorizonDB (Preview) – Cluster Creation, Compute Scaling, Graph DB & AI Pipelines Azure HorizonDB (Preview) unifies relational, vector, graph databases, and AI pipelines into a single auto‑scaling cluster on PostgreSQL, simplifying modern AI workload architecture while offering flexible compute and high availability options.
- Data Modeling and Partitioning in Azure Cosmos DB Data Modeling and Partitioning in Azure Cosmos DB teaches developers to prioritize application read/write patterns over traditional table design, emphasizing embedding related data when it’s naturally contained and choosing partition keys that align with query frequency to optimize performance, scalability, and cost.
- Transparent data encryption in Azure SQL Database now supports AES keys (Public Preview) Azure SQL Database’s public preview of AES key support for transparent data encryption aligns with NSA CNSA 2.0 recommendations, offering a quantum‑resistant alternative to traditional RSA-based protectors by reducing decryption latency and enhancing long-term cryptographic resilience.
- Data API Builder and SQL MCP with Jerry Nixon Data API Builder lets data professionals quickly create secure REST and GraphQL APIs for SQL Server, PostgreSQL, Cosmos DB, and MySQL, integrating security, schema extraction, and governance policies to control access while enabling ad‑hoc queries.
- Protect Azure Cosmos DB with vaulted backups using Azure Backup (public preview) Vaulted backups in the Azure Backup preview isolate Cosmos DB data in a secure, managed vault, providing protection against accidental deletion, ransomware, and compliance needs with policy‑driven automation.
- Build 2026: New Azure Managed Redis Capabilities for AI-Ready Applications Azure Managed Redis at Build 2026 introduces vector search, real-time memory management, and RAG support to enhance AI application responsiveness while adding Entra ID RBAC for secure multi‑team access.
- Announcing new security, maintenance and analytics features for PostgreSQL at Microsoft Build 2026 The announcement at Microsoft Build 2026 introduces a suite of new security, maintenance, and analytics features for Azure Database for PostgreSQL flexible server, including the V6 SKU with NVMe storage, pg_duckdb and pg_ivm extensions for advanced analytics, Defender security assessments, temporal_tables for change tracking
- Microsoft Build 2026: Building agentic apps with Microsoft Fabric and Microsoft Databases Microsoft Build 2026 introduces Rayfin, an open-source SDK and CLI that enables developers to define and deploy enterprise‑grade application backends directly into Fabric, accelerating the transition from prototype to production for agentic AI apps.
- SQL Server Performance Office Hours Episode 62 In Episode 62, Erik Darling answers five community‑submitted questions on indexed view updates, query tuning debates, and free open‑source SQL Server monitoring tools with customizable dashboards.
- Fuzzy String Matching in SQL Server 2025 SQL Server 2025 introduces built-in fuzzy string matching functions—EDIT_DISTANCE, EDIT_DISTANCE_SIMILARITY, JARO_WINKLER_DISTANCE, and JARO_WINKLER_SIMILARITY—that enable developers to assess data quality, detect inconsistencies, and correct human errors directly within T
🛠️ Developer tools
- Generally Available: Microsoft Foundry for Visual Studio Code (June Build 2026 refresh) The generally available launch of the Microsoft Foundry VS Code extension includes a Build 2026 update adding a complete model catalog, interactive playground, and seamless Hosted agent deployment for Azure users.
- Generally Available: Github Copilot modernization GitHub's Copilot modernization agent, now in general availability, automates application assessment and upgrade workflows across large code portfolios through coordinated orchestration.
- Generally Available: Oracle schema conversion to Azure PostgreSQL in Visual Studio Code The new Visual Studio Code extension simplifies migrating Oracle schemas to Azure PostgreSQL by offering a project‑based, code‑editor integrated conversion process.
- Generally Available: GitHub Copilot integration in Schema Designer for the MSSQL extension The GitHub Copilot integration in Schema Designer for the MSSQL extension for Visual Studio Code now offers AI‑assisted suggestions directly within the visual schema design interface, streamlining database modeling tasks.
- Generally Available: Azure DocumentDB Migration Extension in Visual Studio Code The Azure DocumentDB Migration Extension in Visual Studio Code now offers GA support, enabling developers to seamlessly create, execute, and oversee migrations from on‑premises or cloud sources straight within their IDE.
- Generally Available: Data API builder with built-in GitHub Copilot in MSSQL extension The newly GA Data API builder in the MSSQL VSCode extension uses GitHub Copilot to automatically generate REST, GraphQL, and model context code, streamlining backend creation for developers.
- Generally Available: Azure SQL updates for June Azure SQL now offers General Availability features including executing SQL queries directly from Markdown documents and combining Python kernels within SQL Notebooks via the Visual Studio Code extension.
- Public Preview: Observability developer experience in Azure Developer CLI (azd) The preview introduces an observability developer experience within Azure Developer CLI (azd), enabling users to integrate a quality loop into agents built with Microsoft Foundry for enhanced lifecycle management.
- Public Preview: Code-first observability for Foundry Agents in VS Code In the upcoming public preview, Foundry Agents will enable developers to perform code‑first observability directly within Visual Studio Code's Copilot Chat plugin, streamlining the optimization loop entirely inside the IDE.
- Local-first AI Agents in C#: Foundry Local, MEAI, and Microsoft Agent Framework This C# library bridges Foundry Local’s on‑device inference with the Microsoft Extensions.AI and Agent Framework’s IChatClient, enabling seamless local AI app development without REST overhead.
- VS Code 1.123 Adds Agent Session Sync, 1M Context Windows Visual Studio Code 1.123 introduces agent session sync to streamline collaborative workflows and supports up to 1 million tokens in context windows for enhanced productivity.
- The Codebase Is the Prompt: Wolverine, Vertical Slices, and AI-Assisted Development Wolverine's focus on vertical slice architecture compresses code complexity, reducing the number of files an AI coding agent must process per feature and improving prompt efficiency in AI-assisted development.
- GitHub Copilot and tokens: how to keep using AI without burning your budget in three prompts (some personal lessons learned!) GitHub Copilot’s usage‑based billing highlights that optimizing token consumption—by limiting chat history, selective use of autocomplete for small tasks, and choosing appropriate models—can significantly reduce costs while maintaining productivity.
- Episode 429: Getting started with LLM Wikis Episode 429 demonstrates how to set up an LLM wiki using Obsidian, markdown clippings, and AI tools like Claude Code and GitHub Copilot CLI to create a searchable, graph-linked personal knowledge management vault.
- Foundry Toolkit for VS Code at //build: Hosted Agents End-to-End, a Smarter Toolbox, and More Foundry Toolkit’s new end‑to‑end Hosted Agent experience in VS Code streamlines scaffolding, local debugging with an enhanced Agent Inspector, flexible deployment options like ZIP code and BYOI methods, and comprehensive observability features including tracing, continuous evaluation settings, and a dedicated evaluations node for seamless
- Microsoft Ties Majorana 2 Quantum Chip to Its Build AI Push Microsoft links its Majorana 2 quantum chip to the company’s broader Build AI strategy, aiming to accelerate both quantum and artificial intelligence advancements toward a 2029 commercial goal.
- Introducing the New Browser Automation Tool with Toolboxes in Foundry The new Browser Automation Tool in Foundry, powered by Playwright and available as an MCP tool with built‑in authentication, enables AI agents to interact with any website like a human, addressing the last‑mile automation gap for web‑based workflows.
- Keep GitHub Copilot Agent Skills Small and Focused Keeping GitHub Copilot Agent Skills small and focused improves maintainability and ensures each skill remains effective for specific tasks without becoming cluttered with unrelated instructions.
- Build 2026: Microsoft Discovery Hits GA as Redmond Pushes Agentic AI Into R&D Build 2026 reveals that Microsoft has officially launched the GA version of its agentic AI Discovery platform, aimed at integrating advanced AI capabilities directly into scientific and engineering research processes.
- AI alone won’t change your business. The system running it will. Building an effective AI-driven enterprise requires more than powerful models; it demands a comprehensive agent platform that integrates multiple systems, embeds robust governance from design, and enables continuous improvement through human oversight.
- What’s Coming Next in Visual Studio: Our Microsoft Build 2026 Announcements Visual Studio’s upcoming updates focus on integrating AI agents to assist with debugging, profiling, and testing, catching errors before builds start, automating merge conflict resolution, modernizing applications to newer .NET stacks, providing context‑aware skills, allowing users to bring their own keys or models for AI integration,
- AI Tour – Houston 2026 At the sold‑out Houston AI Tour on May 28, 2026, I presented "GitHub Copilot as an AI agent in the developer workflow," demonstrating how Copilot boosts productivity while maintaining quality, and engaged with over 3,600 attendees and multiple Microsoft partners.
- Metered Billing for GitHub Copilot: A Necessary Shift GitHub’s shift to metered billing for Copilot highlights the need for developers to consciously choose models and optimize usage, especially impacting individual creators who previously benefited from subsidized access.
- Terraform MCP Server update The updated Terraform MCP Server v0.5.2 now includes advanced features beyond basic registry lookups, thanks largely to assistance from GitHub Copilot CLI and Claude Opus 4.6, making configuration selection more intuitive for users.
- Claude Code on Microsoft Foundry in VS Code — A Practical Setup Guide (with the gotchas)
- VS Code 1.122 Makes BYOK Easier VS Code 1.122 simplifies Bring Your Own Key (BYOK) model usage by eliminating the need to sign into GitHub, enabling enterprise, offline, and air‑gapped workflows with Anthropic, OpenAI, and Gemini models without authentication hurdles.
🔩 DevOps
- Generally Available: Ingest OTLP signals into Azure Monitor with the OpenTelemetry Collector Azure Monitor now officially supports ingesting OTLP signals via the OpenTelemetry Collector, allowing seamless telemetry data transmission from any OpenTelemetry‑instrumented application or platform with minimal configuration.
- Generally Available: Azure Monitor Service Level Indicators (SLI) Azure Monitor's new SLI and SLO features let teams directly gauge customer application experiences, complementing traditional infrastructure metrics such as CPU usage.
- Generally Available: Simple log alerts in Azure Monitor Simple log alerts in Azure Monitor now offers an intuitive interface that reduces setup time by 40% while maintaining high accuracy for critical log-based notifications.
- Faster az login: introducing --skip-subscription-discovery and targeted --subscription The new Azure CLI flags --skip-subscription-discovery and --subscription let users bypass the time-consuming process of enumerating all subscriptions across many tenants, making az login near-instant for large-scale environments with dozens of tenants or hundreds of subscriptions per tenant.
- Securing CI/CD in an agentic world: Claude Code Github action case Researchers discovered that Anthropic’s Claude Code GitHub Action could leak CI/CD secrets when processing untrusted content, due to the Read tool lacking proper sandboxing, and Anthropic later patched this in version 2.1.128 by restricting access to sensitive system files.
- Azure Repos Permissions Azure Repos Permissions guide explains the four-level security hierarchy—from organization to individual repository—detailing inheritance models and best practices like using explicit deny rules to prevent accidental permission overrides.
- Azure Repos vs GitHub Azure Repos is designed for enterprise centralized development teams with integrated security and CI/CD pipelines, while GitHub excels in open‑source collaboration, community tools, and flexible branching models.
- GitHub Copilot Multi-Repo Instructions: Sharing Skills, Agents, and Conventions Across Repos GitHub Copilot multi-repo instructions enable teams to define and share coding conventions, pitfalls, and best practices across multiple repositories, ensuring that the AI assistant generates code aligned with team standards and automates tasks like PR creation in Azure DevOps.
- DevOps for Microsoft Hosted Agents: From Terraform Apply to Production-Grade Agent Delivery Foundry now supports deploying Hosted Agents directly from source code via a `.zip` package with `code_configuration`, reducing latency for Python or .NET projects while requiring only Project Manager and User roles, complementing the existing container deployment path.
- How Microsoft is migrating repositories to GitHub Migrating to GitHub at Microsoft scale has reduced engineering overhead, enabling over 1,600 repositories and 3,100 developers to transition in six months while preserving critical Azure Boards and Pipelines workflows, unlocking advanced AI capabilities like Copilot Coding Agent.
- GitHub Action for Deploying Hosted Agents This GitHub Action streamlines deploying Hosted Agents in Foundry by using a repeatable workflow with minimal configuration, leveraging existing tooling like Azure CLI and Bash to integrate smoothly into CI/CD processes.
- New Capabilities to Observe Agents in Azure Monitor Azure Monitor now treats agents as first-class observables, offering faster telemetry ingestion, larger event support, a fleet view of monitored agents, deeper debugging insights, and enhanced AI-powered troubleshooting capabilities to improve agent quality evaluation and security compliance.
- Azure DevOps and GitHub: Journeying into the AI Era Azure DevOps and GitHub are deepening their integration with new agentic AI features across planning, coding, review, and security, allowing seamless hybrid usage while enhancing enterprise governance through an agent control plane.
- GitHub Copilot app: The agent-native desktop experience The GitHub Copilot app provides a unified desktop experience that centralizes multiple agent-native development workflows, allowing developers to monitor, manage, and interact with parallel sessions across repositories from a single My Work view.
- Building and Operating a Microsoft Foundry Hosted Agent with GitOps and GitHub Tasks Building a GitOps workflow for a Microsoft Foundry Hosted Agent enables automated, auditable deployments of prompts, tools, and infrastructure changes using GitHub Actions, ensuring consistent environments from development to production.
- Why Choose Pulumi Over Terraform? Pulumi offers stronger programming model tools like first‑class language SDKs, built‑in refactoring workflows with aliases and secrets management, and clearer provider resource definitions, making infrastructure code more like application code while still requiring engineering discipline for drift and provider bugs.
- Building a GitHub Copilot Agent Usage Dashboard The article details building an enterprise-grade GitHub Copilot usage dashboard using OpenTelemetry, Azure Monitor, and Grafana to track agent adoption and workflow popularity with granular insights.
- One prompt, four (sub)agents, and ninety seconds to get a working app Google Antigravity 2.0’s built-in subagents enable rapid creation of an agent team that can simultaneously build a Go backend API, frontend API, and unit tests in just ninety seconds using a single prompt.
- The NEW Bicep Like and Distinct Functions ✨ Azure Bicep's new like() function enables pattern matching for strings, useful for conditional deployments based on naming conventions, while distinct() removes duplicate values from arrays, simplifying configurations and preventing resource redundancy.
- J. Tower: A.I. Workflows - Episode 404 Jonathan "J." Tower, a 12-time Microsoft MVP and .NET Foundation Board member, shares insights on AI workflows and software leadership in this episode of J. Tower: A.I. Workflows.
🧬 Hybrid + multicloud
- Azure Arc: Why you need both WindowsOsUpdateExtension and WindowsPatchExtension Both WindowsOsUpdateExtension and WindowsPatchExtension may appear on Azure Arc‑enabled servers to handle assessment and patching separately, unlike native VMs that use a single handler, ensuring smoother integration with hybrid environments.
- Cert-manager on AKS, the Managed Way cert-manager can be deployed on Azure Kubernetes Service (AKS) clusters using its public preview as an Azure Arc Kubernetes extension, offering managed certificate automation without requiring Arc-enabled infrastructure.
- Sign Into On-Prem Windows Server with Entra ID Using Azure Arc Azure Arc enables on‑premises Windows Server 2025 to authenticate with Entra ID, moving away from Active Directory while using Azure RBAC for centralized security control.
- Build, deploy, and govern sovereign AI with Foundry Local on Azure Local Foundry Local on Azure Local enables customers to build, deploy, and govern sovereign AI applications entirely within their own environment, preserving data residency while leveraging familiar Azure tools for model management and policy enforcement.
- Unlock On-Prem Productivity with Agentic Retrieval in Foundry Local Foundry Local’s Agentic Retrieval in Build 2026 enables on-prem productivity by combining reasoning, agency, and natural user experiences with Azure Arc‑powered edge AI, offering flexible deployment modes, BYOM support, and compliance‑ready solutions for regulated environments.
- Scale On-Prem AI with Foundry Local on Azure Local: Multi-Node Inference and vLLM Support Foundry Local on Azure Local now supports multi-node scheduling, vLLM runtime for high-throughput serving, and an expanded catalog of models optimized formats, enabling scalable, production‑grade on‑premises AI inference in disconnected environments while maintaining Kubernetes‑native, OpenAI‑compatible patterns.
- Introducing GitHub Enterprise Local (Preview): DevOps for Sovereign and Private Cloud Environments GitHub Enterprise Local lets government, defense, finance, and critical infrastructure firms run GitHub’s full enterprise developer platform on their own sovereign cloud infrastructure with zero internet dependency, preserving data control while maintaining a familiar GitHub experience.
🎭 Identity
- What’s New in Entra Global Secure Access Client v2.31.125 The updated Entra Global Secure Access Client v2.31.125 adds a Network Status feature that gives administrators real-time visibility into connection health and simplifies troubleshooting processes.
- Step-by-Step guide to Microsoft Entra Agent ID – Part 03 – Create Attribute-Driven Conditional Access Policy for Agent Identities This guide demonstrates how to use Attribute‑driven Conditional Access policies in Microsoft Entra Agent ID to automatically grant or block access for agent identities based on custom security attributes like "type," enabling scalable and dynamic governance without manual policy updates.
- Patch your Domain Controllers now CVE‑2026‑41089 is an unauthenticated, wormable remote code execution vulnerability in Windows Netlogon that can grant SYSTEM privileges on domain controllers, allowing attackers to extract sensitive data, manipulate Group Policy, and spread ransomware across the network.
- Build AI agents for production with secure identities from day one Microsoft Entra Agent ID provides a consistent identity foundation for deploying AI agents in production, ensuring secure identities, access controls, and auditability from day one.
🔌 Integration
- Generally Available: Azure API Management Premium v2 and Standard v2 now support wildcard custom hostnames. Azure API Management Premium v2 now supports wildcard custom hostnames, simplifying the management of multiple subdomains and reducing operational complexity for growing API estates.
- Generally Available: Azure Event Grid - MQTT v5 Subscription Identifier Azure Event Grid now offers generally available support for MQTT v5 subscription identifiers, allowing apps to route incoming events instantly based on the included unique identifier.
- Generally Available: Azure API Management Premium v2 now supports multiple custom domains Azure API Management Premium v2 now supports hosting multiple custom domains within a single instance, allowing organizations to brand APIs distinctly while simplifying management.
- Generally Available: Azure API Management now supports AI gateway capabilities for Anthropic and Vertex AI models. Azure API Management's new AI gateway capabilities let platform teams centrally secure, govern, and monitor interactions with Anthropic and Vertex AI models in production applications.
- Generally Available: Azure Logic Apps MCP Server. Azure Logic Apps MCP Server simplifies AI agent workflows by eliminating the need for custom API layers, reducing development time and operational overhead.
- Generally Available: Azure API Management adds support for Agent-to-Agent (A2A) APIs Azure API Management now offers General Availability support for Agent-to-Agent (A2A) APIs, enabling better governance and security for inter-agent communications within agentic application ecosystems.
- Generally Available: Azure API Management now supports content safety controls for MCP and A2A APIs. Azure API Management's new content safety controls help organizations enforce uniform security and compliance standards when building agentic applications and multi-channel conversational (MCP) tools, reducing the need for piecemeal solutions.
- Generally Available: Azure API Management workspaces now support the built-in gateway Azure has simplified API Management by allowing built-in gateways in workspaces, eliminating the need for separate Premium-tier deployments and reducing costs and complexity.
- Generally Available: Azure API Center now provides a data plane MCP server for enterprise-wide discovery of APIs and AI assets. Azure API Center's new data plane MCP server streamlines enterprise-wide discovery of APIs and AI assets by centralizing connections to multiple MCP servers into a single searchable interface.
- Generally Available: Azure API Center now supports agent registration, agent assessment, and Git-based synchronization. Azure API Center's new features let organizations register, assess, and synchronize AI agent APIs via Git, streamlining governance and reuse for multi-team deployments.
- Public Preview: APIM Support for Foundry Models in Azure AI Search The public preview of APIM Support for Foundry Models in Azure AI Search enables enterprise teams to seamlessly integrate Foundry models with API Management, streamlining large-scale retrieval‑augmented generation (RAG) pipelines.
- Public Preview: Azure API Management introduces a Unified Model API for multi-model AI applications. Azure API Management's Unified Model API standardizes interactions across various LLM providers, simplifying development and enhancing flexibility for multi-model AI applications.
- Public Preview: Azure API Management now supports token metrics for all token types Azure API Management's public preview introduces token metrics for a wider range of token types—including cached, reasoning, and thinking tokens—to help organizations better monitor advanced AI model consumption.
- Public Preview: Azure Functions now includes managed connectors to integrate with 1400+ of your favorite systems Azure Functions gains managed connectors for over 1,400 popular services, allowing seamless integration into workflows without custom code.
- Public Preview: Azure Logic Apps Codeful Workflows with Logic Apps Standard SDK. Azure Logic Apps introduces a Codeful preview that lets developers write workflows in code, combining custom logic with Azure's managed orchestration and extensive connector ecosystem.
- Public Preview: Azure Logic Apps Automation SKU for agentic business process automation. Azure Logic Apps introduces an Automation SKU designed specifically for agentic business process automation, streamlining the integration of workflows, APIs, and AI agents with a single, production‑ready solution.
- Public Preview: Azure Logic Apps now enables developers to invoke Microsoft Foundry Agents directly from workflows. The preview reveals that Azure Logic Apps will let developers directly invoke Microsoft Foundry Agents, streamlining the workflow between AI agent development and automation of business processes.
- Public Preview: Knowledge as a Service with Azure Logic Apps. Azure Logic Apps introduces Knowledge as a Service, simplifying Retrieval-Augmented Generation by handling ingestion, chunking, embeddings, vector storage, and retrieval layer configuration out of the box.
- Giving Developers Claude Code with Azure API Management and Claude Models in Microsoft Foundry By routing Claude Code requests through Azure API Management with Entra ID authentication, organizations can centrally manage per‑developer token usage, enforce rate limits, and track costs in Foundry without exposing Anthropic API keys or dealing with billing sprawl.
- Build Enterprise-Ready Agents with Microsoft IQ and Oracle AI Database@Azure — now with Oracle MCP The new Oracle Managed Database MCP (Model Context Protocol) connector for Azure simplifies AI integration by reducing CDC latency to under five seconds, enabling real‑time context updates for tables up to 100 million rows while maintaining seamless interoperability with existing workflows.
- Azure Event Grid: Powering IoT and Event-Driven Applications at Scale Azure Event Grid introduces MQTT v5 Subscription Identifiers for faster message routing and cleaner application logic, supports 1MB messages to reduce architectural complexity, and offers preview autoscaling capabilities to automatically adjust capacity based on real‑time workload demand.
- Write Logic Apps in C#: introducing the Logic Apps Standard SDK The Logic Apps Standard SDK lets .NET developers author workflows in C# with type safety, IntelliSense, and familiar tooling while retaining all runtime capabilities of the existing visual designer.
- New AI gateway capabilities in Azure API Management Azure API Management introduces a Unified Model API that standardizes multi-model access through a single OpenAI Chat Completions interface, simplifying observability, security, and failover management across diverse AI providers.
- 🎉 Automation just became a team sport. Meet Azure Logic Apps Automation. Azure Logic Apps Automation introduces a low‑barrier, SaaS‑style experience that lets non‑engineers create production‑ready workflow automations in minutes while preserving Azure’s enterprise security, governance, and scalability.
- 📢 Announcing Knowledge as a Service for Azure Logic Apps Knowledge as a Service for Azure Logic Apps simplifies turning documents into an instant, accurate knowledge base without building or maintaining complex RAG pipelines, streamlining access to institutional information for AI agents and workflows.
- Better Together: Build Agents in Microsoft Foundry, Automate them with Azure Logic Apps This integration lets developers build sophisticated agentic AI agents in Microsoft Foundry and automatically trigger them using Azure Logic Apps’ extensive connector ecosystem, enabling seamless business process automation without custom code.
- Azure Connector Namespaces: managed integration for any Azure compute Azure Connector Namespaces offers fully managed integration for any Azure compute, handling authentication, credential rotation, polling, webhook delivery, retries, throttling, and error handling through a consistent programming model with SDKs or HTTP, filling the gap between custom API clients and workflow engines.
- Hosted MCP Servers in Connector Namespace (Preview) Connector Namespace’s hosted MCP servers let developers quickly deploy secure, managed remote MCP instances with built‑in authentication, scaling, observability, and tool discovery, offloading operational overhead from infrastructure to platform maintenance.
- What's new in Azure Logic Apps at Microsoft Build 2026 Azure Logic Apps at Build 2026 introduces public preview features like Azure Logic Apps Automation, which simplifies enterprise-grade AI-powered automation with built‑in AI assistance and natural language generation, alongside General Availability of the MCP Server for integrating existing workflows as AI tools.
- Announcing managed connectors for Azure Functions (Preview) Azure Functions now offers managed connectors from its extensive 1,400+ ecosystem—like Office 365 and Teams—in preview, enabling native trigger integration and typed SDKs for seamless event‑driven workflows without additional authentication or HTTP handling.
- What's new in Azure API Management at Microsoft Build 2026 Azure API Management at Build 2026 introduces AI gateway enhancements, an Azure API Center that centralizes discovery and governance of APIs, agents, MCP tools, and AI assets, plus support for JSON‑RPC Agent‑to‑Agent APIs with content safety controls to manage emerging agentic workloads.
- MCP Test Console and Git Repository synch in Azure API Center Azure API Center now integrates MCP Test Console and Git repository synchronization, enabling developers to test tools directly within the portal and automatically keep MCP server definitions up to date from version-controlled repositories.
- More Control, Less Overhead: Custom Domain Upgrades in Azure API Management v2 Azure API Management Premium v2 now allows multiple custom domains within a single instance, enabling organizations to configure distinct hostnames for different endpoints and audiences while reducing operational complexity and the need for separate API Management instances.
- Azure API Center Introduces a Data Plane MCP Server for Enterprise-Wide API and AI Asset Discovery The Azure API Center launches a data plane MCP server that centralizes discovery of enterprise APIs, tools, and AI assets, reducing manual configuration and integration complexity as the MCP ecosystem expands.
- Find what you need, faster: Azure API Center now supports custom metadata filtering Azure API Center now lets enterprises filter catalog assets using custom metadata like environments, business units, and compliance tiers, streamlining developer productivity by aligning search with existing organizational classifications.
- Build a profiling harness before you benchmark Building a profiling harness before benchmarking clarifies performance issues by isolating code paths, reducing noise from production complexity, and ensuring measurements reflect real-world conditions accurately.
💡 Internet of Things
- Generally Available: Azure Sphere OS version 26.06 is now available Azure Sphere OS 26.06 GA introduces enhanced security patches and IoT device management tools, prioritizing edge reliability for manufacturers deploying over 10,000 units.
🎓 Learning and Certifications
- Recent and upcoming Microsoft exam changes – Sunday 7 June 2026 The article details the retirement of the AI-900 Microsoft Azure AI Fundamentals exam on June 30, 2026, along with upcoming changes to other exams.
- Build AI skills at Microsoft AI Skills Fest Microsoft AI Skills Fest offers a free digital event where partners can develop in-demand AI skills, earn certification through the Frontier Transformation Engineer badge, and learn to build innovative solutions across platforms like 365 Copilot and Fabric.
- AI-103 vs AI-200 – Which Azure AI Certification Is Right for You? AI‑200 focuses on building cloud‑native infrastructure for AI back‑ends—containers, serverless functions, event pipelines, and observability—making it ideal for developers who already handle Azure services but need expertise in scaling AI solutions, while AI‑103 is geared toward creating AI applications and agents using
- From campus to career: 3 practical steps for students Leverage free resources like the Microsoft AI Skills Fest, GitHub Education pack, and Student Ambassadors program to develop in-demand AI skills, showcase projects with proof‑of‑work on platforms like GitHub, and build leadership experience that sets you apart in a competitive job market.
- 345 - Reflections from Microsoft Build 2026 This episode highlights standout announcements from Microsoft Build 2026, including Windows 11’s Intelligent Terminal and Run dialog enhancements, GitHub Enterprise Local, Azure Logic Apps Automation, and the debut of Project Solara alongside new AI models and execution containers.
- Microsoft AZ-204 certification is Being Replaced by Microsoft AI-200 Certification The new AI-200 certification focuses on containerised compute, AI‑enabled data services, event‑driven pipelines with Azure OpenAI and managed identity security, replacing AZ-204 as Azure developers must now build production‑ready AI solutions.
- Azure Developer Associate Renewal – Study Guide – 2026 The Azure Developer Associate renewal guide for 2026 highlights essential study resources and the upcoming July 31 deadline, emphasizing key vault storage, App Config keys, Application Insights monitoring, Container Apps secrets management, and Azure Functions triggers.
- DevOps Engineer Expert Renewal – Study Guide – 2026 The guide compiles essential Azure DevOps resources, including GitHub permissions and ADO agents, to help candidates efficiently prepare for the 2026 renewal certification.
⚖️ Management and Governance
- Public Preview: Azure Infrastructure Resiliency Manager Azure Infrastructure Resiliency Manager streamlines resiliency planning by integrating Availability Zones, Azure Advisor recommendations, and Chaos Engineering into a single goal-oriented dashboard.
- Public Preview: Azure Policy Coverage for Model Router in Foundry Models Azure Policy's new coverage for Model Router in Foundry Models enables organizations to centrally govern and enforce routing standards, ensuring model selections comply with security, regulatory, and operational policies.
- From cloud adoption to value realisation From cloud adoption to value realisation focuses on measuring tangible benefits like reduced recovery time and retired legacy hosting contracts after deploying Azure services.
- Azure Monitor Health Model (Preview): What's New! Azure Monitor's new Health Model (Preview) consolidates telemetry, architecture, and business data into a single actionable health state, reducing alert noise and enabling proactive operations with flexible discovery options like Application Insights, Azure Resource Graph, and Service Groups.
- AI Insights for Database Monitoring: A DBA’s Perspective This article offers a DBA’s firsthand account of how AI can uncover critical database issues like repeated backup failures and log reuse states, providing actionable insights for proactive monitoring.
- Is 94% of your syslog just noise? Now you can filter it out before ingestion. Multi-stage transformations in Azure Monitor let you filter and aggregate syslog data on the agent side, cutting unnecessary volume by up to 94% before it’s ingested, while also reducing costs and improving query performance.
- When Telemetry Volume Gets Real: Azure Monitor pipeline’s Performance Story! Azure Monitor pipeline can sustain over 200,000 Syslog messages per second on an 8‑core node, processing about 17 billion events or 20 TB daily while using only ~2.8 GB of memory, demonstrating impressive linear scalability and efficient resource utilization
- Azure Policy Best Practices Azure Policy Best Practices emphasizes defining policies at the Root Management Group level for scalability, using targeted exemptions for temporary needs, bundling related rules in Initiative Definitions to simplify governance, and implementing a phased rollout with Audit → Modify → Deny effects to balance visibility and enforcement.
- Why Microsoft 365 knowledge hubs fail without governance and content quality A well‑governed Microsoft 365 knowledge hub delivers value only when its information is accurate, trusted, and adheres to clear ownership standards; otherwise, even sophisticated tools like Copilot will surface low‑quality or outdated content.
- Is Your Monitoring Actually Working? What's New in Monitoring Coverage Monitoring Coverage in Azure Monitor now includes data flow status and at‑scale alert enablement, helping teams quickly identify missing monitoring configurations and ensure real‑time alert delivery across virtual machines and AKS clusters.
- What’s new in Observability at Build 2026 Azure Build 2026 introduces AI‑focused agent observability in Azure Monitor, integrating OpenTelemetry for portable signals, new Agent Observability features with fleet views and automated evaluations, and an expanded Copilot Observability agent that enhances natural language investigation across AKS, Application Insights, and Foundry AI
- Azure Monitor Copilot Observability Agent: What’s new at Build The updated Azure Monitor Copilot Observability Agent at Build 2026 introduces expanded investigation scenarios, integration with Microsoft Foundry AI Agents for cross‑scenario signal correlation, and enhanced capabilities in Application Insights, AKS clusters, Activity Logs, and future support for multiple Application Insights resources and Azure Service Health events
- Any source. Any destination. Ready for AI-era. Azure Observability at Build 2026 introduces an AI-native OpenTelemetry standard, bursty traffic scaling to billions of events per day, unified governance for AI and platform telemetry via DCRs, multi‑stage noise control, and comprehensive Monitoring Coverage to ensure complete signal reliability across data journeys.
- Infrastructure as Code for AI: Building and Deploying Microsoft Hosted Agents with Terraform Terraform automates provisioning of Azure infrastructure for Microsoft Hosted Agents, ensuring repeatability, governance, and scalability while leaving runtime management to the Managed Agent Service.
- Azure Governance Best Practices Azure Governance Best Practices outlines a four-pillared framework—resource organization, security & access control, compliance & guardrails, and cost management—to establish scalable, enterprise‑wide cloud governance across multiple subscriptions and resource groups.
- WSUS Is Deprecated. That's Not Really the Story. While WSUS deprecation sparks questions about legacy management, many customers are also exploring modern alternatives like Azure Update Manager and transitioning to Linux environments.
🚌 Migration
- Generally Available: Azure Files assessments are now available worldwide using Azure Migrate Azure Migrate's new global support for assessing SMB and NFS file shares enables customers to gain comprehensive, data-driven insights into their entire file share infrastructure across all regions.
- Announcing: Confidential Live Migration for Intel TDX confidential VMs in Azure Confidential Live Migration in Azure enables seamless relocation of Intel TDX confidential VMs during updates without exposing data, enhancing both security and uptime.
- Azure Storage Mover Azure Storage Mover is a fully managed hybrid migration service that centralizes file and folder transfers from on-premises environments to Azure Storage, offering high performance, metadata preservation, and cost efficiency compared to alternatives like Azure Data Box and AzCopy.
- Announcing Confidential Live Migration in Azure Azure’s new Confidential Live Migration lets Intel® TDX VMs move to updated infrastructure with minimal interruption while preserving memory isolation and encryption during the transition.
- File share migrations simplified with Azure Copilot Migration Agent Azure Copilot Migration Agent simplifies file share migrations by integrating discovery, assessment, planning, and execution into a single guided workflow in Azure Migrate, reducing operational friction and speeding up migration to Azure Files.
- Architecting Resilient AI Solutions on Microsoft Azure with Regions and Availability Zones Azure’s Regions and Availability Zones enable architects to build AI solutions with multi‑fault tolerance, low latency access across continents, and seamless scaling to meet growing workloads.
🌐 Networking
- Generally Available: Managed virtual network for evaluations in Microsoft Foundry With the new managed virtual network feature, users can securely evaluate cloud workloads in isolated environments without exposing internal IP ranges.
- Private Preview: Guest RDMA on Azure Boost Guest RDMA on Azure Boost offers up to 50% faster data transfer between virtual machines by leveraging direct memory access, ideal for high-performance computing and gaming workloads.
- MANA support for Existing VM SKUs:Why Now Is the Right Time to Update Linux on your SAP on Azure VMs Microsoft’s new MANA support for existing Azure VM SKUs, particularly M-series and Mv3 families, lets SAP customers on Linux stay current with newer platform capabilities like improved compatibility, performance boosts, and security updates without immediate infrastructure overhauls.
- Designing Cloud Landing Zones by Traffic Flow: A Defence‑in‑Depth, DMZ‑First Architecture Designing Cloud Landing Zones by Traffic Flow introduces a defence‑in‑depth, DMZ‑first architecture that classifies enterprise traffic—Internet inbound, outbound, east‑west, and hybrid connectivity—to enforce distinct inspection requirements at well‑defined trust boundaries, using a hub‑and‑spoke model
- Announcing Preview of Guest RDMA for Azure Boost Azure Boost’s Guest RDMA preview in the UK South region enables virtual machines to achieve kernel‑bypass, ultra-low latency networking with up to 100Gb/s throughput, ideal for AI/ML training, distributed storage, and HPC workloads by offloading data movement directly to network interface cards
- Inspecting AKS East-West Traffic with Next-Generation Firewalls: Solving a Problem Kubernetes… The article details how organizations can apply Next-Generation Firewalls to inspect East-West traffic in Azure Kubernetes Service using innovative architectures like Application Gateway, Istio Gateways, or Private Link Service, despite the inherent challenges posed by Azure CNI Overlay's encapsulated pod-to-pod communication.
🔐 Security
- Defender for Endpoint EDR Updates Move to Microsoft Update Channel Microsoft is shifting Defender for Endpoint EDR updates from monthly cumulative patches to the Windows Update channel, requiring verification that Intune update rings permit other Microsoft products to receive these changes.
- Enhancements to Device Status API & Logged-In User Email in Endpoint DLP The Endpoint DLP Device Status API now provides real‑time visibility into device health, configuration, and policy sync state, enabling admins to instantly identify logged‑in users during incidents and streamline compliance reporting with live dashboards instead of manual Excel exports.
- Extend Microsoft Purview data protection to AWS Bedrock agents for cross-cloud AI governance Microsoft Purview extends data protection to AWS Bedrock agents, enabling consistent cross-cloud AI governance with real-time DLP policies for hybrid workloads.
- Updating the taxonomy of failure modes in agentic AI systems: What a year of red teaming taught us The updated taxonomy of failure modes in agentic AI systems introduces seven new categories—such as Agentic Supply Chain Compromise and Computer Use Agent Visual Attacks—and incorporates real-world red team findings to enhance security mitigation strategies.
- Microsoft Defender for Cloud Customer Newsletter The new integration of Defender for Cloud into the Defender portal streamlines cloud security posture management and threat protection, offering customers a unified experience with real-time reporting capabilities.
- Run Global Secure Access with confidence: Introducing the GSA Operations Guide The new Microsoft Entra Global Secure Access operations guide provides actionable, alert‑first procedures and role matrices to help teams confidently manage day‑two operational challenges after deployment.
- What’s new in Microsoft Sentinel: May 2026 Microsoft Sentinel's May 2026 update introduces unified role‑based access controls with row‑level scoping for granular permissions across Sentinel and Defender, expands connector catalog to over 400 integrations via Codeless Connector Framework, and previews Agent 365 for AI agent telemetry in the data lake.
- Preinstall to persistence: Inside the Red Hat npm Miasma credential-stealing campaign The Red Hat npm Miasma campaign used a hijacked CI/CD pipeline to publish trojanized packages with authentic signatures, which executed a heavily obfuscated script during installation to steal credentials from GitHub, cloud services, and developer systems before propagating through maintainer packages.
- The end of patching era for containers: Microsoft Defender for Cloud expands hardened image support Microsoft Defender for Cloud now supports scanning and managing vulnerabilities in popular hardened container image types like Chainguard, Minimus, and Docker Hardened Images, centralizing security oversight while reducing attack surface through minimal, transparently built images.
- Microsoft Build 2026: Securing code, agents, and models across the development lifecycle The new Microsoft Security multi-model agentic scanning harness (MDASH) integrates AI agents with Defender and GitHub Code Security to discover, validate, and prove exploitability across codebases using a configurable mix of high-performance and cost-effective models.
📦 Storage
- Generally Available: File share centric management model (Microsoft.FileShares) for Azure Files The newly released File Share centric management model in Azure Files lets users directly manage file shares as top-level resources, simplifying security and access controls for NFS 4.1 shares stored on SSD.
- Public Preview: Secure, Modern Access to Azure Files on macOS with MS Entra ID Azure Files introduces a public preview that lets macOS users securely access file shares via MS Entra ID, enhancing cross-platform identity-based authentication and modern workflow integration.
- Azure Storage Mover vs AzCopy Azure Storage Mover excels for enterprise-scale hybrid migrations with centralized orchestration in the cloud, while AzCopy is ideal for ad‑hoc, script‑driven data transfers requiring minimal overhead and developer control.
- Your GPv1 Storage Accounts have a migration deadline, or a “do nothing” option that’ll quietly cost you money Migrating Azure GPv1 or legacy Blob storage accounts by October 13, 2026, is mandatory; otherwise, Microsoft will auto‑migrate them to GPv2 with potentially higher transaction costs depending on workload intensity.
💻 Virtual Desktop Infrastructure
- Webinar: Certificate Automation in Practice – A Technical Deep Dive The webinar explores practical techniques for automating certificate management in Windows Server environments, ensuring seamless renewal of critical TLS certificates without disrupting services like VPNs, DirectAccess, and IIS.
- Teams Remote App/ Cloud App optimization for Windows 365 and Azure Virtual Desktop now GA The new media engine in Teams Remote App and Cloud Apps for Windows 365 enhances audio, video quality, reliability, and security while simplifying updates without frequent infrastructure changes.
- Adaptive data protection with context-based redirections in Windows 365, now in public preview Windows 365’s public preview of context‑based redirections lets admins tailor clipboard, drive, printer, and USB access in real time based on device compliance, user groups, and network conditions to enhance security while maintaining productivity.
- Azure Virtual Desktop supports greater application and identity functionality with latest updates Azure Virtual Desktop now generally supports external identities with FSLogix profile management, SAML domainless IdP configuration in preview, and US Government cloud access for external users, enhancing cross‑platform deployment options.