Azure Weekly

Issue 572

26^th July 2026

A last minute holiday meant that I couldn't do the newsletter last week, so it's a bumper edition. As ever, if you want a better reading experience for such a long newsletter (275 articles!), have a look at the web edition.

My highlights are as follows:

• Claude Opus 5 is available today in Microsoft Foundry by Amar Badal - Claude Opus 5 in Microsoft Foundry supports long-running, error-resilient agents for complex enterprise workflows with integrated advanced vision capabilities for handling visual and textual documents on a secure platform.
• Announcing the Open-Source Release of ML Video Codec (MLVC) by Naba Kumar - MLVC replaces traditional video primitives with end-to-end neural compression that runs on commodity NPUs and is currently powering video calls in Microsoft Teams. Available on GitHub under a MIT license.
• From Live Signals to Smarter Risk: Building a Company Brain on Microsoft Fabric by Dori Rabin - Microsoft Fabric converts live maritime AIS signals into a real-time agentic risk intelligence system that integrates vessel data, geospatial boundaries, policy documents, and master datasets for complex insurance question answering.
• Azure Kubernetes Service (AKS) on Bare Metal by Thomas Maurer - Azure Kubernetes Service on bare metal enables organizations to run Kubernetes in on-premises, edge, or sovereign environments while maintaining Azure’s simplicity and consistency.
• Use GitHub Agentic Workflows to test Bicep deployments by Leo Visser - This article shows how to use GitHub Agentic Workflows and Azure tools such as Bicep, PowerShell, and the Entsoe Transparency Platform to automatically test Azure deployments for sustainability, optimize region selection based on energy mix and latency, and ensure compliance with best practices.
• Introducing Azure Front Door edge actions - Bringing secure, programmable logic to the edge by Akhil Karmalkar - Azure Front Door now provides edge actions enabling developers to run lightweight JavaScript functions globally for real-time personalization, security, and resiliency without added backend complexity.

Finally, something we've been working on for the last few years: Rx.NET v7.0 Now Available by Ian Griffiths - Rx.NET v7.0 reduces self-contained deployment size by up to 90 MB and separates UI framework support into distinct NuGet packages, addressing a long-standing issue with Windows-specific projects. This is a big win if you want to include Rx as part of your cloud projects.

🤖 AI

• The AI Agent Lifecycle: A Simple Guide
  The AI Agent Lifecycle guide outlines four stages-design, build, test, and monitor-with tools like Foundry Model Catalog, Azure AI Content Safety, and the Foundry Agent Service to ensure safety, compliance, and continuous improvement.
• Creating Autonomous Teams Agents Using OpenClaw, MCP, and Azure Container Apps
  OpenClaw enables autonomous coding agents to automate development cycles in chat platforms, raising security and control concerns for production.
• Brady Gaster: Squad AI Agent - Episode 410
  Brady Gaster, GitHub’s Principal PM Architect, introduces Squad AI Agent, an open-source framework enabling developers to orchestrate collaborative teams of specialized AI agents using GitHub Copilot.
• How to Cut GenAI and Agent Token Spend Without Cutting Capability
  This guide optimizes GenAI and agent token usage by tracking total task run consumption through a token control plane, separating token reduction from rate limiting for precise savings reporting.
• Agentic Data Systems: Why AI Agents Need More Than a Vector Database
  Agentic data systems need to advance past vector databases to handle speculative high-volume queries, maintain structured memory for persistent tasks, and incorporate verification tools due to limitations in managing agents' branching, retries, and coordination.
• Agent Memory Is Architecture, Not a Feature
  Designing agent memory by defining storage locations, access controls, expiration policies, and retrieval methods is crucial to prevent both forgetting vital information and creating dangerous data swamps, as stated in the article.
• Building AI Agents from Zero to Production
  Building AI Agents from Zero to Production is a seven-lesson open-source course using Azure's MAF and Foundry stack to guide developers through designing, developing, evaluating, deploying, securing, tooling, and orchestrating multi-agent systems for enterprise production environments.
• Finding Security Vulnerabilities using AI with Sami Laiho
  Sami Laiho discusses how AI models like Anthropic’s Mythos and Microsoft’s MDASH are transforming vulnerability discovery, prompting an arms race to patch flaws before malicious actors exploit them with large language models.
• How Clustering can be Used as a Preprocessing Technique for Classification in Azure Machine Learning
  Azure Machine Learning demonstrates how clustering can preprocess data, enabling the creation of more accurate classification models for specific subsets within large datasets.
• Building a Fully Managed Multi-Agent Pipeline with Microsoft Foundry
  The workshop shows how to use the Foundry Toolkit in VS Code to create a "Resume → Job Fit Evaluator" multi-agent workflow with four specialized AI agents using fan-out and fan-in patterns, simplifying complex business processes.
• Answers You Can Trust: Grounding Enterprise Agents with Foundry IQ
  Foundry IQ enables enterprise AI agents to ground their answers in an organization's own knowledge base, ensuring traceability, permission control, and compliance for critical business queries.
• Meet the IQ's: How Microsoft is Creating Context-Aware AI
  Microsoft's IQ Platform integrates Work IQ, Fabric IQ, and Foundry IQ to offer contextual awareness through enterprise data and semantic understanding of work processes.
• Tool Calling Governance Is the New Agent Security Boundary
  Tool calling governance is the new agent security boundary as it focuses on managing tool owners, purposes, schemas, and approval rules to mitigate operational risks.
• Choose the right speech and voice model
  Choosing the right speech and voice model in Microsoft Foundry depends on whether you need conversation control, managed agent stacks, transcription artifacts, speaker diarization, high-quality narration, or multilingual translation services.
• HolmesGPT + Azure AI Foundry: A Quick Kubernetes Troubleshooting Demo
  HolmesGPT combined with Azure AI Foundry demonstrates a rapid Kubernetes troubleshooting workflow, where an AI agent diagnoses scheduling issues in under five minutes using minimal setup requirements.
• Agents can learn with Memory in Microsoft Foundry Agent Service
  Microsoft Foundry Agent Service introduces procedural memory to enhance AI agent learning from past experiences, boosting efficiency without labeled training data.
• Hello World: Your First Azure OpenAI Call in C#
  This introductory C# sample demonstrates a minimal console app integrating Azure OpenAI via the Microsoft Agent Framework, showing prompt handling, model responses, and token usage.
• From Chatbot to Agent: Adding Tool Calling
  The tutorial shows how function invocation lets models autonomously choose and loop through tools such as date/time retrieval, weather lookup, and calculations, turning chatbots into agents with notable boosts in token usage per tool call.
• Building Agents in Production with Toolbox, Skills, and Tool Search
  Join a live Discord round table on July 22, 2026, to discuss how Microsoft’s Toolbox, Skills, and Tool Search capabilities can simplify integrating and managing tools for AI agents in production.
• MCP Turns Agent Tools Into an Enterprise Boundary
  MCP standardizes how AI agents interact with enterprise systems, enabling secure, auditable tool access and governance through centralized servers rather than custom integrations.
• When RAG Fails, Treat Retrieval Like a Production System
  Treating retrieval like a production system with quality gates, lifecycle management, monitoring, and incident response is crucial when RAG fails to prevent weak context from harming the model.
• Round Table: Building Browser-Capable Agents with the Browser Automation Tool
  Join a 40-minute interactive Discord round table on July 22 at 2:30 PM BST (7:00 PM IST) to discuss how the Browser Automation tool in Microsoft Foundry helps agents perform real browser workflows, share your web automation challenges, and suggest improvements.
• Azure Content Understanding: Classify and Route to Different Analyzers
  Azure AI Content Understanding automatically classifies mixed-document PDFs like invoices, bank statements, and loan applications for processing with separate analyzers using the .NET SDK.
• Public Preview: Protect sensitive generative AI telemetry in Application Insights and Microsoft Foundry
  Azure Monitor's public preview adds a GenAIContent table for fine-grained access control of sensitive generative AI telemetry data.
• Important LLM benchmarks to know about
  Understanding key LLM benchmarks such as ARC-AGI-2 for abstract reasoning, Terminal-Bench 2.1 for agentic coding, and ExploitBench for exploitation helps clarify each model's strengths in general intelligence, coding tasks, or security challenges.
• Beyond Benchmarks: The Lifecycle of Measuring Agentic Quality in AI Content Management
  The article describes a comprehensive evaluation lifecycle for assessing agentic quality in AI content management within OneDrive and SharePoint, focusing on continuous multidimensional quality measurement through rapid inner-loop validation and real-world customer feedback.
• Choosing the Right Starting Point for Enterprise AI Agents with Copilot Studio and Microsoft Foundry
  Choosing between Copilot Studio and Foundry depends on team expertise; low-code builders prefer Copilot Studio for quick business use, while developers opt for Foundry's code-first method for intricate custom AI apps.
• Grounding Copilot Studio Agents with Azure AI Search and Foundry IQ
  The article showcases five retrieval patterns for an HR knowledge agent using Copilot Studio, Azure AI Search, and Foundry IQ, catering to various user needs with a shared knowledge base.
• Designing the AI Agent Execution Boundary: What Agents Should Never Be Allowed to Do
  Designing the AI Agent Execution Boundary involves setting strict limits to prevent unauthorized permission expansion, identity misuse, destructive actions without approval, and change management bypasses for safety and accountability.
• Design, test, and ship Foundry hosted agents from a canvas in GitHub Copilot App
  The Foundry Agent Canvas in GitHub Copilot App streamlines designing, testing, and deploying hosted agents with a unified visual interface that reduces friction during rapid experimentation.
• Introducing MAI-Image-2.5 Pro and MAI-Voice-2 Flash in Microsoft Foundry
  MAI-Image-2.5 Pro targets professional creative workflows needing high visual fidelity and creative control, whereas MAI-Voice-2 Flash focuses on low-latency voice applications for real-time interactions in 15 languages.
• AT&T and Microsoft scale trillion-token workloads with Microsoft Foundry and AMD
  AT&T used Azure’s Foundry Managed Compute and AMD GPUs to train Open Telco (OTel) 2.0 for telecom-specific AI at trillion-token scale with optimized cost, performance, and flexibility.
• 1 million documents to 300+agents: Building an enterprise-scale Microsoft 365 Copilot connector
  The new enterprise-scale Microsoft 365 Copilot connector integrates 1 million documents across 100,000+ users and 300+ agents, showcasing design choices for security, governance, schema decisions, operational patterns, and retrieval quality to enhance internal knowledge accessibility in Copilot.
• 🏆 Agents League: Celebrating the Builders Who Made Agents Battle for Glory
  The Agents League hackathon showcased winners like Afterlogin: The Hunt, which used one engine for both a SOC simulation and training game, demonstrating how multi-agent orchestration can serve diverse user experiences with high polish.
• For the first time, real-time transcription goes multilingual
  Multilingual Post-Stream Refinement automatically detects spoken languages in real-time streams, enabling simultaneous transcription of multiple languages with low latency and improved accuracy.
• Post-Stream Refinement is now generally available in Microsoft Foundry
  Post-Stream Refinement in Azure AI Speech’s Foundry now offers production-grade speaker diarization, phrase list support for domain vocabulary, and coverage across 19 locales in 22 regions, maintaining real-time streaming while improving final transcript accuracy.
• Microsoft Foundry hosted agents: lessons learned building a production multi-agent service
  Building a production multi-agent service on Azure AI Foundry’s hosted agents teaches key lessons on choosing the right compute model and protocol for bursty or scheduled workloads, preserving session state across idle periods, and leveraging dedicated Entra identities per agent to limit blast radius while benefiting from built-in telemetry.
• AfterLogin: We Turned Forgotten-Account Cleanup into a Haunted-House Game with Microsoft Foundry
  AfterLogin transforms forgotten-account cleanup into a haunted-house game with AI providing evidence for triaging accounts, solving the issue of distinguishing dormant from active accounts.
• Announcing the Open-Source Release of ML Video Codec (MLVC)
  MLVC replaces traditional video primitives with end-to-end neural compression that runs on commodity NPUs and is currently powering video calls in Microsoft Teams.
• Claude Opus 5 is available today in Microsoft Foundry
  Claude Opus 5 in Microsoft Foundry supports long-running, error-resilient agents for complex enterprise workflows with integrated advanced vision capabilities for handling visual and textual documents on a secure platform.
• Available today: Anthropic Claude Opus 5 in Microsoft 365 Copilot
  Claude Opus 5 boosts Microsoft 365 Copilot with advanced features for complex coding, professional work, and long-term reasoning in Word, Excel, PowerPoint, Chat, Copilot Cowork, and Copilot Studio while preserving user control.

🔎 Analytics

• Build and Optimize a Data Lakehouse for Unified Data Intelligence
  Building a data lakehouse on Azure offers unified governance, low-latency analytics, and scalable architecture for BI dashboards, AI agents, and decision-support systems without duplicating data.
• Generally Available: Azure Databricks SQL Serverless in UK West
  Azure Databricks SQL Serverless in UK West offers instant compute and automated scaling, simplifying analytics workloads for customers in that region starting July 13th, 2026.
• Where Should a Data Quality Check Live? Gates, Controls, and Quarantine
  Data quality checks should be implemented early to prevent errors effectively.
• Azure Databricks delivers proven business value
  The Forrester study reveals Azure Databricks provides a 331% return on investment with $58.1 million net present value, recouping costs in under six months via improved productivity, lower infrastructure costs, better data resilience, and saved DBA time.
• From Live Signals to Smarter Risk: Building a Company Brain on Microsoft Fabric
  Microsoft Fabric converts live maritime AIS signals into a real-time agentic risk intelligence system that integrates vessel data, geospatial boundaries, policy documents, and master datasets for complex insurance question answering.
• Building an Apache Arrow Parser for Power BI’s ExecuteDaxQueries Endpoint
  This article explains how to create an Apache Arrow parser for Power BI's ExecuteDaxQueries endpoint to improve native data type processing in Power Query.
• Fabric Workspace Monitoring: What It Actually Shows You
  Fabric Workspace Monitoring streams execution logs into a dedicated Eventhouse, enabling custom KQL queries and automated alerts via Activator for pipelines, semantic models, and databases.
• Using Fabric Operations Agents And Workspace Monitoring With Power BI
  The Operations Agent monitors Power BI for DAX query errors, alerting on issues while providing user, date/time, error message, and resource IDs via Workspace Monitoring integration.
• Fabric apps, featuring the Field Technician App template
  The Fabric apps collection introduces a Field Technician App template that streamlines job registration, task assignment, and role-based dashboards for dispatchers and technicians within real-time intelligence solutions.
• SQL Server vs Fabric Warehouse: T-SQL gotchas nobody warns you about
  Fabric Warehouse’s T-SQL compatibility hides critical differences-such as the absence of IDENTITY keys, non-enforced primary/foreign constraints, restricted ALTER TABLE options, limited data types (e.g., no XML or geography), and unsupported features like recursive CTEs and BULK LOAD-that can catch developers off guard when migrating from SQL Server.
• Power BI Performance in 2026: A Practical Guide to Faster Semantic Models and Reports
  This updated guide explains that Power BI performance results from complex factors including data source, storage mode, semantic model, and refresh processes, emphasizing architectural choices over superficial tweaks.
• Think Twice Before Enabling Fabric’s Inbound Network Protection
  Enabling Fabric’s inbound network protection can degrade or disable critical features like on-premises data gateways, SQL Server mirroring, pipelines, Data Warehouse operations, Eventhouse real-time analytics, Purview scanning, and Power BI governance without clear documentation.

🖥️ Compute

• Azure In-Place Upgrade: Windows Server 2012–2025
  Azure's in-place upgrade enables safe transition of unsupported Windows Server 2012-2025 VMs to supported versions, reducing security risks and compliance issues while maintaining roles, settings, and data.
• Your H100s on Azure won't do distributed training. Here's probably why.
  Azure H100 training may stall if the wrong SKU is chosen or ND VMs aren't placed in a single placement group, disabling necessary InfiniBand connectivity.
• Azure Cobalt: Workload-Aware Power Management for More Efficient Datacenters
  Azure Cobalt introduces per-virtual-machine power monitoring and capping to improve datacenter efficiency by up to 20%, intelligently managing non-critical workloads.
• Ragnarök for Windows Server: A Practitioner's Guide to Extended Security Updates
  Ragnarök for Windows Server details how Extended Security Updates (ESU) provide up to 10 additional years of security patches, crucial for legacy servers without direct support.
• Public Preview: Azure Functions Support for PowerShell 7.6
  Azure Functions now offers public preview support for PowerShell 7.6, enabling developers to build and deploy serverless applications using the latest PowerShell version directly on Azure.
• Generally Available: Azure Functions support for Python 3.14
  Azure Functions now provides general availability for Python 3.14, offering improved security, extended support, and smooth compatibility for developing and deploying serverless apps on Linux.
• Trusted Launch for virtual machines on Windows Server Insider Preview
  Trusted Launch for Gen 2 VMs in Windows Server Insider Preview enhances security with Secure Boot, vTPM, and state protection against boot-level attacks, managed via PowerShell.
• Connecting Microsoft Discovery App to Azure HPC with Azure NetApp Files and CycleCloud
  The article explains how the Microsoft Discovery app connects with Azure HPC using Azure NetApp Files and CycleCloud, enabling agentic AI workflows to interact seamlessly with traditional POSIX-based file systems and scheduler orchestration in HPC environments.
• Azure capacity constraints in Europe: why North Europe is feeling it most
  Azure's capacity constraints in North Europe are severe due to energy grid strain and policy changes, leading to frequent VM creation failures and strict quota management affecting IaaS and PaaS services.

🚢 Containers

• Orchestrate Azure Container Apps Jobs with Apache Airflow
  Orchestrate Azure Container Apps Jobs with Apache Airflow lets you run complex, dependent pipelines-like parallel batch jobs and conditional retries-while leveraging ACA’s serverless scaling, all without provisioning a separate scheduler.
• AKS managed Gateway API blocks the ALB controller
  The article explains how AKS's managed Gateway API conflicts with the ALB controller, causing version issues that keep Application Gateway for Containers stalled and Front Door to return 504 errors until custom Helm setups are applied.
• Azure Kubernetes Service (AKS) on Bare Metal
  Azure Kubernetes Service on bare metal enables organizations to run Kubernetes in on-premises, edge, or sovereign environments while maintaining Azure’s simplicity and consistency.
• How Microsoft 365 built a platform engineering layer on AKS to ship faster at global scale
  Microsoft 365's COSMIC platform streamlines provisioning, security, and compliance on Azure Kubernetes Service, cutting operational complexity for quicker innovation.
• AKS Node Disruption Policy is Now in Preview: Control When Your Nodes Get Reimaged
  The new AKS Node Disruption Policy preview lets users schedule node reimaging during off-hours, giving greater control over cluster maintenance and minimizing business disruption.
• Generally Available: Encryption in Transit for Azure Files NFS Shares in Azure Kubernetes Service (AKS)
  Azure Kubernetes Service now generally offers Encryption in Transit for Azure Files NFS shares, ensuring TLS-encrypted data transfer between AKS workloads and the file storage.
• Microsoft Agent Framework Multi-Agent Workflow Architecture for Automated Kubernetes Assessments
  The Microsoft Agent Framework automates Kubernetes assessment generation to ensure objective, scalable, and resilient tests through concept structuring, deterministic validation, real-cluster behavioral checks, and self-healing repair loops.
• How to Install and Configure the NVIDIA GPU Operator on Kubernetes
  The NVIDIA GPU Operator simplifies Kubernetes integration with GPUs through automated driver installation, toolkit setup, device plugin configuration, and resource discovery, necessitating hardware visibility checks and version compatibility for smooth operation.
• Hybrid Logic Apps on RKE2: a self-managed cluster with MetalLB
  Deploying Azure Logic Apps on an RKE2 self-managed Kubernetes cluster requires configuring MetalLB for ingress IP handling node inotify limits manually setting up CoreDNS naming conventions and using the correct distribution flag key steps differing from OpenShift deployments.
• Introducing Kubernetes-Native Policy Validation with CEL and VAP in Azure Policy
  Azure Policy now integrates Kubernetes-native validation with CEL and VAP for lower latency, stronger fail-close behavior, while keeping existing governance features.

🗄️ Databases

• Lessons Learned #543: Evaluating MultiSubnetFailover with Azure SQL Database
  Evaluating MultiSubnetFailover for Azure SQL Database showed no clear improvement in connection recovery during failovers, as the service's routing model differs from SQL Server high availability configurations.
• Lessons Learned #544: How to Detect INT Identity Exhaustion Before Inserts Fail.
  This article shows how to use a SQL query with sys.identity_columns to monitor INT IDENTITY columns and avoid exhaustion by detecting near-zero remaining values, enabling timely migration to BIGINT before insert failures occur.
• A Programmatic Interface to the Query Hint Recommendation Tool
  The new PowerShell script automates SQL Server Management Studio's Query Hint Recommendation Tool to identify optimal query hints for fully automated performance tuning.
• PostgreSQL on Azure: Two services, one future-proofed ecosystem
  Azure provides two PostgreSQL services-one for general workloads via Azure Database for PostgreSQL's open-source features and reliability, and a new Azure HorizonDB optimized for high-performance AI applications with sub-millisecond latency and elastic scaling.
• Generally Available: SQL Migration to SQL Server on Azure Virtual Machines in Azure Architecture
  Azure Arc now provides GA support for migrating SQL Server workloads to Azure Virtual Machines with minimal downtime via online backup and log shipping.
• Security Update for SQL Server 2016 SP3 Azure Connect Feature Pack
  The Security Update for SQL Server 2016 SP3 Azure Connect Feature Pack resolves multiple vulnerabilities, notably a critical Denial of Service issue (CVE-2026-54118), and can be downloaded from the Microsoft Download Center and Update Catalog.
• Security Update for SQL Server 2022 RTM
  The Security Update for SQL Server 2022 RTM GDR fixes several vulnerabilities, notably a critical DoS issue (CVE-2026-54118), available for download from the Microsoft Download Center or Update Catalog.
• Public Preview: Azure SQL updates for mid-July 2026
  Azure SQL's July 2026 public preview introduces Shortcut Configuration for customizing keyboard shortcuts in Visual Studio Code for tools like Quick Queries, the Results Grid, and the Query Editor.
• June 2026 Recap: Azure Database for PostgreSQL
  Azure Database for PostgreSQL introduced Chaos Studio Workspaces in public preview, enabling automated zone-down failover simulations to test high availability configurations.
• Generally Available: Microsoft Defender security assessments for Azure Database for PostgreSQL Flexible Server
  The General Availability of Microsoft Defender CSPM assessments for Azure Database for PostgreSQL Flexible Server enables continuous automated evaluation of security configurations to ensure compliance and reduce vulnerabilities with detailed remediation guidance.
• Connect an Azure Function to Azure SQL database Using a Managed Identity
  Using Azure Functions' Managed Identity simplifies secure access to Azure SQL Database by eliminating passwords, centralizing identity management via Azure Active Directory, applying least privilege permissions, and enabling automatic credential rotation.
• Monitoring and using pg_repack in Azure Database for PostgreSQL flexible server
  Monitoring pg_repack in Azure Database for PostgreSQL flexible server verifies active operation tracks progress phases and assesses resource impact on large tables.
• CDK Global modernizes automotive CRM on Azure SQL Managed Instance
  CDK Global upgraded its automotive CRM on Azure SQL Managed Instance to enhance efficiency, data utilization, and rapid market adaptation while maintaining continuity for dealership operations.
• Can You Really Downgrade from Enterprise to Standard and Save Thousands? Here's What Happened
  Downgrading from SQL Server Enterprise to Standard Edition saved the customer $89,016 while maintaining performance for Analysis Services and Power BI Report Server by removing Enterprise-only features.
• Two additional controls with Automatic Plan Correction enabled
  The undocumented sp_configure_automatic_tuning procedure provides detailed query-level control for Automatic Plan Correction, enabling exclusion from monitoring and enhanced performance checks for diverse workloads.
• Identify SQL Server EOL & ESU Costs with azqr
  The azqr sql-esu plugin identifies near-end-of-life SQL Server instances and estimates Extended Security Update costs to guide modernization to Azure SQL Managed Instance.
• Regex-based dynamic data masking in Azure SQL Database (preview)
  Azure SQL Database now offers regex-based dynamic data masking to precisely obfuscate sensitive string fields such as emails and phone numbers while retaining necessary information.
• [Video] Office Hours: Long Answers Edition
  In this Office Hours video, Brent Ozar answers top-voted SQL Server questions on topics including HADR_sync_commit delays, indexing strategies for multi-row tables, transitioning isolation levels, comparing MSSQL and PostgreSQL, assessing 2025 readiness, TempDb bottlenecks, and filtered indexes vs statistics.
• SQL Agent Job Automated Change Capture Process
  This solution provides a way to track historical changes of SQL Agent jobs with a dedicated table and stored procedure, addressing the lack of native audit capabilities in SQL Server.
• Meet Darling: Free, Headless Fleet Performance Monitoring for SQL Server
  Darling provides free headless fleet performance monitoring for SQL Server with 36 collectors, a self-hosted PostgreSQL database, real-time alerts via email/webhook, and an MCP server for AI-assisted configuration management.
• mssql-python v1.12.0: Standalone ODBC package, bulk copy fixes
  mssql-python v1.12.0 introduces a standalone mssql-python-odbc package for better space management and independent updates, fixing bulk copy functionality to respect connection timeouts and enable writing to custom CLR UDT columns.
• TLS Certificate Pinning and Best Practices in Azure Database for PostgreSQL
  Azure Database for PostgreSQL recommends using a CA-based TLS trust model instead of certificate pinning to avoid connection failures during automatic certificate rotations while maintaining security.
• SQL Data Sync Retirement: Migration Insights and Modern Alternatives
  SQL Data Sync will retire on September 30, 2027, so organizations should adopt Azure Data Factory for better data synchronization.

🛠️ Developer tools

• OpenAPI Code Generation with Corvus: Typed HTTP Clients
  Corvus' OpenAPI code generation creates typed HTTP clients with schema validation, reducing allocations 12-17 times versus Kiota and providing exhaustive response handling at compile time for .NET projects.
• SkiaSharp for .NET
  SkiaSharp 4.0, co-maintained by Uno Platform and Microsoft’s .NET team, provides GPU-accelerated cross-platform rendering for .NET MAUI, WinUI 3, and the Uno Platform, ensuring consistent high-performance graphics across all supported devices.
• Visual Studio Takes Aim at Copilot Billing Shock
  Visual Studio’s recent update introduces features like enhanced AI-assistance for developers, improved security tools, and various quality-of-life tweaks alongside clearer Copilot billing visibility.
• The Visual Studio Dev/Test Benefit: Freedom to Build, Test, and Experiment in Azure
  Azure Dev/Test pricing within a Visual Studio Subscription offers discounted non-production Azure resources, enabling developers to experiment freely without worrying about cloud costs.
• OpenAPI Code Generation with Corvus: Server Stubs
  Corvus generates server stubs for ASP.NET Core APIs, handling parsing, validation, and error responses directly from OpenAPI specs so handlers contain only business logic.
• JetBrains Resharper for Visual Studio Code adds C# debugging support
  JetBrains’ Resharper for Visual Studio Code now supports C# debugging, offering an alternative to the licensed C# Dev Kit extension and addressing limitations for forks like AWS Kiro and Cursor.
• Built-in Agent Skills Bring .NET and Azure Expertise into Visual Studio
  Visual Studio now includes built-in .NET and Azure Agent Skills to help developers customize agentic workflows for tasks like API development and performance analysis.
• OpenAPI Code Generation with Corvus: End-to-End
  Corvus generates both client and server code from a single OpenAPI spec, ensuring tight contract enforcement and seamless round-trip communication between them with built-in authentication handling.
• Customer Offerings: Solution Optimization for GitHub Copilot
  Solution Optimization for GitHub Copilot helps customers gain visibility into usage, set practical cost controls, and improve agent quality to achieve more useful outcomes per interaction.
• Pick, manage, and get the most from your models
  Visual Studio allows pinning favorite language models for comparison and monitoring real-time context window usage.
• Announcing: DevUI Agent Inspector for Microsoft Agent Framework
  The DevUI Agent Inspector offers developers a local debugging tool to visualize, pause, replay, and edit agent prompt executions in real time.
• .NET 11 Preview 6 Roundup: ASP.NET Core, MAUI, C#, EF Core and SDK Updates
  .NET 11 Preview 6 introduces significant updates to ASP.NET Core, MAUI, and C#, including enhanced async validation, C# union types, improved UI controls, optimized EF Core queries, expanded testing tools, Native AOT support, and new container features.
• Announcing: Agent Channel for Microsoft Agent Framework multi-agent communication
  The new Agent Channel in the public preview of the Microsoft Agent Framework enables multi-agent communication with typed messages, eliminating the need for custom routing code.
• Announcing: Episodic procedural memory in Microsoft Agent Framework
  The new episodic procedural memory feature in the public preview of the Microsoft Agent Framework allows AI agents to retain and reuse successful task patterns, enhancing efficiency for repetitive operations.
• Announcing: CodeAct pattern and Hyperlight containers in Microsoft Agent Framework
  The new CodeAct pattern reduces end-to-end latency by optimizing sequential tool calls within a single code block.
• Announcing: Tracing for Microsoft Agent Framework (Python and .NET)
  The new tracing feature for the Microsoft Agent Framework lets developers monitor agent reasoning, tool calls, and model interactions in real-time using Python or .NET.
• Announcing: Multi-agent orchestration SDK with the Microsoft Foundry Agent Framework in C# and Python
  The new Multi-agent Orchestration SDK from Microsoft Foundry unifies pattern management in C# and Python, streamlining development with consistent abstractions for multi-agent systems.
• Announcing: Agent Harness in Microsoft Agent Framework
  Agent Harness in the Microsoft Agent Framework is now generally available for direct production use with backward compatibility.
• Announcing: Multi-agent orchestration patterns including Magentic in Microsoft Agent Framework
  The new General Availability of the Multi-agent Orchestration Patterns in the Microsoft Agent Framework includes the dynamic Magnetic pattern, showing a 38% improvement on GAIA and strong results on AssistantBench and WebArena benchmarks.
• Announcing: GitHub Copilot and Claude Code connectors in Microsoft Agent Framework
  The new connectors let developers create .NET and Python agents that seamlessly offload coding tasks to GitHub Copilot or Claude Code, eliminating the need for custom adapter development.
• OpenAPI Code Generation with Corvus: Streaming
  Corvus enhances OpenAPI code generation for streaming responses, supporting Server-Sent Events and NDJSON to handle LLM completions, real-time feeds, and event-driven UIs with typed, validated documents on both client and server sides.
• No Mono in Microsoft’s cross-platform MAUI framework on .NET 11 – and a quick hands-on
  Microsoft's .NET 11 MAUI now runs exclusively on CoreCLR, improving iOS and Mac Catalyst performance while removing Mono support, with Android maintaining about 10% startup speed.
• Visual Studio Administrator? Join our Private Marketplace Preview!
  Private Marketplace in Visual Studio lets organizations host, configure, and control private extensions for enhanced security, compliance, and governance within the familiar IDE workflow.
• VS Code 1.129 Introduces Agent Host and Experimental Agents Window Editor
  VS Code 1.129 introduces an Agent Host to run AI sessions separately and an experimental Agents Window Editor for reviewing generated code changes directly within the IDE.
• Copilot Usage-Based Billing Gets a Token Dashboard
  The token dashboard for Copilot’s usage-based billing tracks how many extra tokens Agent Skills consume, with Visual Studio’s integration currently disabled by default.
• OpenAPI Code Generation with Corvus: Authentication
  Corvus generates config for standard .NET auth patterns like IHttpClientFactory and Azure.Identity without adding a custom auth layer, ensuring smooth integration with OAuth2.
• Rx.NET v7.0 Now Available
  Rx.NET v7.0 reduces self-contained deployment size by up to 90 MB and separates UI framework support into distinct NuGet packages, addressing a long-standing issue with Windows-specific projects.
• The cost of saying yes has changed
  Using AI agents to quickly generate small code patches lets teams validate feature requests in minutes instead of days, revealing hidden costs and helping decide whether a change is truly worth the effort.
• INFO: Anthropic Claude or Open AI Codex as coding agents for GitHub Copilot
  You can now use Anthropic’s Claude and OpenAI’s Codex as coding agents within GitHub Copilot for free, accessible via the GitHub web UI, mobile app, and VS Code with shared context.
• How to Migrate Your .NET 8 and 9 Apps to .NET 10 with Visual Studio and GitHub Copilot
  Visual Studio and GitHub Copilot streamline the .NET 8 to .NET 10 migration, making it a simple version bump with automated NuGet package updates as apps approach end-of-support on November 10, 2026.
• OpenAPI Code Generation with Corvus: Callbacks, Webhooks and Links
  Corvus improves OpenAPI code generation by offering distinct tools for creating webhook receivers and clients, ensuring typed validation and easy role reversal between API servers and subscriber apps.
• GitHub Agentic Workflows vs Microsoft Scout vs Copilot App: Which Tool to Use When
  GitHub Agentic Workflows automate repository tasks unattended, while Microsoft Scout serves as an always-on personal agent for productivity tasks in 365 and on your desktop, with GitHub Copilot app focusing on supervised coding sessions.
• Copilot AI Billing Shock Met with Meters, Caps and Token-Saving Tools
  GitHub introduces metered billing, caps, and token-saving tools for Copilot, alongside Microsoft enhancing Visual Studio and VS Code with features to monitor and minimize AI development costs.
• Which Pilot is Which?
  Which Pilot is Which? explains the differences among GitHub Copilot, Copilot Studio, the GitHub Copilot App, and Microsoft 365 Copilot.
• Microsoft Agent Framework for .NET (part 2): How to use IChatClient with Azure Foundry and Ollama in .NET
  The article demonstrates configuring IChatClient from the Microsoft Agent Framework with Azure Foundry for messaging infrastructure and Ollama for on-device LLM integration in .NET applications.
• OpenAPI Code Generation with Corvus: Filtering
  Corvus' new filtering options enable developers to generate focused OpenAPI clients with only necessary operations, enhancing navigation and compile times while maintaining code alignment with service dependencies.
• How to build interactive experiences with canvases
  Canvases enable developers to interact directly with visual information for real-time collaboration through customizable interfaces.
• Visual Studio 2026 Gives Copilot Built-In Skills -- and Makes Them Prove Their Worth
  Visual Studio 2026 introduces built-in Copilot skills authored by experts, which remain disabled until performance tests confirm they add sufficient value to justify extra computational cost.
• OpenAPI Code Generation with Corvus: Testing
  Testing focuses on business logic rather than the automatically generated serialization, parameter encoding, schema validation, or response parsing; use unit tests with isolated handlers and integration tests using ASP.NET Core’s WebApplicationFactory to verify behavior and error handling.
• Microsoft Agent Framework Makeover: Claws, Loops and Harnesses
  The new Microsoft Agent Framework Harness simplifies development by integrating essential components like loops, planning, memory, context management, and safety controls directly into the framework.
• Copilot vs. raw API access: What are you actually paying for?
  Copilot offers integrated development tools and organizational controls alongside model access, while raw API access is better for building custom systems where you manage prompts, retrieval, routing, security, and billing.
• Middleware in Microsoft Agent Framework
  Middleware in the Microsoft Agent Framework allows developers to control token usage, improve security, boost performance, and add custom logic by intercepting and altering API calls between users and AI agents.
• VS Code 1.130 Expands Agent Host and Review Tools
  VS Code 1.130 introduces enhanced agent host capabilities and new review tools, enabling better collaboration on multi-file changes and improved visibility in chats and terminals.
• Beyond the Prompt: Building Enterprise Solutions with AI & Specification-Driven Design
  Adopting Specification-Driven Design with disciplined markdown specs and modular architecture allows developers to create secure, compliant enterprise AI solutions such as a zero-trust healthcare risk-assessment engine while preserving system integrity.
• Your Entire Agentic AI Workflow, Now Inside VS Code: New Course Available
  The Foundry Toolkit for Visual Studio Code streamlines agentic AI workflows by integrating model selection, agent development, and testing directly within the IDE, boosting productivity.
• Visual Studio Insiders Further Refines Copilot Usage Tracking
  Visual Studio 18.9 simplifies reaching the monthly Copilot plan limits and provides clearer cost details within the model picker to help developers manage usage-based billing concerns.

🔩 DevOps

• Optimizing GitHub Copilot Cost in the Usage-Based Billing Era
  Optimizing GitHub Copilot cost involves selecting appropriate models, limiting token usage through context windows and session length, implementing budget controls, and adopting a tiered pricing strategy to balance productivity with financial management.
• Distributing Bicep Modules with Token-Based Access Using ACR
  Bicep 0.45.6 adds ociEnabled for token-based module distribution to Azure Container Registry with ORAS and Docker credentials, simplifying management for MSPs and platform teams.
• I let AI build my API. Here’s what happened (building an app with an AI LLM, part four)
  Grant Fritchey uses AI tools to develop a secure REST API linking a PostgreSQL database on Azure with an upcoming WordPress frontend while documenting successes and challenges.
• Exporting Every Availability Test URL from a Subscription with PowerShell
  The PowerShell script extracts and deduplicates ping target URLs from Azure Application Insights availability tests for easy CSV export, aiding preparation for new URL ping testing costs.
• Load testing Copilot Studio agents with Locust and Azure Load Testing
  Load testing Copilot Studio agents with Locust and Azure Load Testing shows successful operation under heavy use, with one turn taking up to 8.68 seconds to reply and 163.71 seconds to finish, emphasizing the need to measure both scaling capacity and component performance for reliable conversational AI deployments.
• Public Preview: Advanced platform metrics in Azure Monitor
  Advanced platform metrics in Azure Monitor now offer detailed insights into performance, resource health, and operational trends for all supported Azure services starting July 15, 2026.
• GitHub for Beginners: Your roadmap to mastering the GitHub essentials
  GitHub for Beginners provides a concise roadmap covering version control basics, secure account setup with two-factor authentication, and essential Git commands to help readers confidently navigate repositories, collaborate on projects, and contribute to open source.
• Monitoring GitHub Copilot at Scale with OpenTelemetry (OTLP) and Grafana
  This guide shows how to use OpenTelemetry and Grafana to monitor GitHub Copilot at scale for tracking AI adoption, latency, token usage, and costs.
• Adding Observability with OpenTelemetry and Application Insights
  This guide demonstrates integrating OpenTelemetry with Azure Application Insights to monitor AI call behavior, manually initiate custom activities for enhanced telemetry, and analyze token usage metrics using KQL.
• Logging Prompts and Responses for Auditing and Debugging
  Turning on sensitive data capture logs full prompts and model responses alongside telemetry enables robust debugging and compliance but raises privacy risks and logging costs that must be carefully managed.
• The AI-Native Software Development Lifecycle: A Practical Guide for Developers, Leads, and Architects
  The AI-Native SLD focuses on managing challenges such as defining accurate problems, providing reliable context, controlling agent behavior, and ensuring safety and correctness as generative AI speeds up code production.
• Use GitHub Agentic Workflows to test Bicep deployments
  This article shows how to use GitHub Agentic Workflows and Azure tools such as Bicep, PowerShell, and the Entsoe Transparency Platform to automatically test Azure deployments for sustainability, optimize region selection based on energy mix and latency, and ensure compliance with best practices.
• Production-ready, AI-assisted Development and Workflows
  The Pluralsight course covers ensuring AI-generated code is secure, resilient, observable, and cost-effective by integrating AI coding assistants with human oversight.
• Keep the gains: performance regression testing without fooling yourself
  Keep only benchmarks that protect critical hot paths, use baseline/diff comparisons with clear thresholds, and avoid relying on noisy shared runners to reliably catch performance regressions without false positives or negatives.
• Developing with AI Exposes Your Bottlenecks
  Developing with AI can shift bottlenecks from coding to other process stages like code review or testing if those steps remain unchanged, according to the Theory of Constraints.
• Logging & OpenTelemetry in MAF
  Logging and OpenTelemetry in the Multi-Agent Framework improve debugging, performance monitoring, and audit trails, enhancing observability and system efficiency.
• Activity Source in Microsoft Agent Framework
  The integration of OpenTelemetry into the Microsoft Agent Framework enhances observability through unified task tracking, automatic metric instrumentation, and seamless trace shipping to Azure AI Foundry for comprehensive performance analysis.
• July Patches for Azure DevOps Server
  July’s patches for Azure DevOps Server include updates to the main product and a specific patch for the 2022.2 version, both aimed at enhancing security and fixing critical issues.
• How to Run Coding Agents Safely Inside CI/CD Pipelines
  Implement a safe CI/CD architecture by treating agent output as untrusted, using ephemeral sandboxes, granting read-only repo access, and separating the agent sandbox from a clean verification runner.

🧬 Hybrid + multicloud

• Patch the Kernel, Skip the Reboot: The Case for Arc-Enabled Hotpatching on Windows Server 2025
  Windows Server 2025 adds Arc-enabled hotpatching to apply security updates in memory, cutting planned restarts from monthly to roughly four per year with no additional costs.
• When to Keep AI On-Prem: Data Gravity, Latency, Sovereignty, and Cost as Architecture Inputs
  When deciding whether to keep AI on-prem, organizations must consider data gravity, latency requirements, regulatory constraints, and cost implications for optimal component placement.
• Generally Available: Expanding Azure Arc SQL Migration with SQL Server on Azure Virtual Machines
  Azure Arc now extends SQL migration capabilities to include SQL Server on Azure Virtual Machines for seamless transitions to Azure SQL Managed Instance and native Azure infrastructures via a unified process.
• Lessons Learned #546: Maintaining a Local Azure Resource Inventory
  The PowerShell script maintains a local Azure SQL logical servers inventory in JSON, comparing it with Azure Resource Graph to validate and update only new, missing, or changed resources, reducing repeated API requests.
• Empowering the Mission at the Edge with Azure Local
  Azure Local empowers mission-critical workloads at the edge by extending Azure services to customer-controlled locations, ensuring consistent security and governance even in Denied, Disrupted, Intermittent, or Limited environments.
• Azure Local expands SAN capabilities with iSCSI support
  Azure Local's new iSCSI SAN support expands flexibility by enabling the use of existing IP-based storage networks with a consistent Azure management experience across hybrid and fully disaggregated deployment models.

🎭 Identity

• Use the new Entra SOC Identity Responder Admin Role
  The new Entra SOC Identity Responder Admin Role streamlines incident response by centralizing authority to manage and remediate compromised user accounts, enhancing security team efficiency.
• Microsoft Entra ID security updates: Passkeys are the default authentication method in Entra ID
  Microsoft will make passkeys the default authentication method in Entra ID, phasing out SMS and voice methods by September 1, 2026, to enhance security.
• Microsoft are set to retire SMS & Voice authentication methods
  Microsoft plans to phase out default SMS and voice multi-factor authentication in favor of phishing-resistant Passkeys, with a detailed timeline and guidance on auditing affected tenants.
• AI agents are everywhere. Are your access controls ready?
  Nine out of ten AI agent deployments experience unmanaged sprawl, leading to unexpected thousands of agents without governance visibility.
• Now available: Admin control for SSO prompts in Windows
  The July 2026 Windows update allows IT admins to automatically approve SSO permissions via a registry setting, simplifying sign-in for enterprise devices without affecting personal accounts.
• Why identity is the Copilot unlock
  Identity foundation issues such as lacking multifactor authentication, conditional access policies, and least-privilege models in Azure Active Directory hinder healthcare organizations from fully deploying Copilot despite having purchased licenses.
• Microsoft Entra Suite hands-on tour of identity and network access protections
  Microsoft Entra Suite automatically removes outdated permissions upon role changes, integrates lifecycle workflows with HR systems for seamless access updates, and offers granular control over AI usage via verified identity checks and prompt injection protection.
• MCP Enterprise Authorization Is Here — What Entra and App Service Can Do Today
  Enterprise Managed Authorization in MCP enables centralized access control, streamlining authentication with an identity provider's JWT grant and reducing per-server prompts for organizations managing hundreds of employees.
• Least privilege for AI agents: Identity, access, and tool binding
  Implementing least privilege for AI agents involves managing their identities, applying role-based access controls, tightly scoping permissions, and safely binding tools to prevent unauthorized data access and potential privilege escalation.
• Conditional Access in the Real World – Break-Glass Accounts, Legacy Auth, and Named Locations
  This article explains how to set up break-glass accounts as a critical safety net for managing emergency access in Conditional Access policies within Microsoft 365 Entra ID environments.
• Understanding Microsoft Entra ID Group Membership Caching and Azure SQL Authentication Timing
  Understanding how Microsoft Entra ID groups cache membership explains temporary Azure SQL authentication failures after a user joins a group, which resolve as cached data expires.
• Designing the AI Agent Identity Lifecycle in Microsoft Entra
  Designing the AI Agent Identity Lifecycle in Microsoft Entra focuses on governing identities from creation to retirement, covering classification, blueprint selection by trust boundaries, and ongoing monitoring to avoid identity sprawl and maintain compliance.
• Plan your Azure AD B2C migration with the Migration Policy Analyzer
  The Azure AD B2C Migration Policy Analyzer assesses custom policies for migration readiness to Microsoft Entra External ID, categorizing features accordingly.
• Secure AI, web, and private apps with Zero Trust
  Zero Trust security now includes AI agents, data protection, and private apps with new public preview features for improved visibility and policy enforcement.
• Entra Chat Podcast MCP Server
  The new Entra Chat Podcast MCP Server converts the podcast into a searchable knowledge base, enabling direct queries from tools like Claude Desktop and Cursor with zero infrastructure overhead.
• Microsoft Entra ID enhances security of branded sign-ins
  Microsoft Entra ID will retire custom CSS positioning properties on October 26, 2026, to reduce phishing risks by preventing deceptive page layouts.
• 352 - So long, SMS and voice authentication
  Microsoft will phase out native SMS and voice authentication in Entra ID, replacing it with passkeys by September 1, 2026, and retiring all Microsoft-provided telecom services on February 1, 2027.
• Entra Internet Access TLS Inspection and Always On VPN: Troubleshooting SSTP Connection Failures
  Entra Internet Access offers TLS inspection and Always On VPN to secure user connections, with troubleshooting tips for common SSTP failure scenarios.
• Not All MFA Is Equal – Phishing Resistant MFA and Authentication Strengths in Entra ID
  Phishing-resistant MFA methods like FIDO2 passkeys and Windows Hello provide stronger protection than SMS or push notifications, which are vulnerable to SIM swapping and real-time phishing attacks.
• Governance Is the New Bottleneck: What Agent 365 Means for Admins ?
  Agent 365 adds governance tools to manage and secure AI agents in Microsoft 365, mitigating risks from uncontrolled agent proliferation.
• Understanding Microsoft 365 Copilot Risk Surface and Mitigations
  Understanding the two-layer risk modelaccess to Copilot itself and the data it surfaceshelps organizations apply Zero Trust principles to mitigate exposure from identity, device, app, and data governance perspectives.
• Modernize SAP Identity Management with Microsoft Entra
  Modernizing SAP Identity Management with Microsoft Entra provides flexible provisioning, custom attributes, account discovery, secure OAuth 2.0 communication, and seamless integration for streamlined identity governance across SAP and non-SAP systems.
• Microsoft Entra to allow passwordless password changes
  Starting in October 2026, users of passwordless authentication in Microsoft Entra can update their My Sign-Ins passwords using passkeys or Windows Hello without needing supplemental strong passwords.
• What's New in Microsoft Entra: July 2026
  The July 2026 update to Microsoft Entra introduces features like automatic tenant backup and recovery, direct admin assignment for external users via email, BYOD support for the Global Secure Access client on non-domain-joined devices, enhanced Kerberos key rotation for reliability, and domainless SAML federation that simplifies access for external users without matching email domains.

🔌 Integration

• A message broker is only as reliable as the layer you stopped at
  Azure Service Bus provides layered reliability with regional redundancy, zone-level protection, and geo-replication to customize protection against different failure magnitudes.
• From Multi-Model Chaos to a Governed AI Gateway: Cost Optimization on Azure
  Azure's Governed AI Gateway unifies multiple LLMs through one secure endpoint, applying cost and security controls to cut expenses by up to 30% and ease compliance.
• Using Azure API Management to obtain Anthropic Messages APIs logs
  Azure API Management supports logging and tracking token consumption for Anthropic Messages APIs under the AI Gateway feature in v2 tiers, enabling effective monitoring of both streaming and non-streaming requests.
• Announcing Public Preview of Autoscale for Azure Event Grid Namespaces
  Autoscale for Azure Event Grid namespaces adjusts throughput units in real time to match workload demands, preventing manual provisioning and throttling while maintaining cost limits.
• Putting Azure API Management in Front of Your AI Model
  Azure API Management centralizes key management, rate limiting, and diagnostic logging for shared AI models with separate subscription keys and queryable token usage for cost management.
• Combining Agents with API Management
  The video shows how to integrate a tool-calling agent with API Management using W3C correlation properties and custom headers to track model calls and token usage in one interaction.
• Are Azure Logic Apps really low-code or no-code?
  Azure Logic Apps excels at orchestrating workflows with low-code ease, ideal for simple processes though not optimal for complex data transformations or advanced error handling.
• Azure API Center: Link APIM for automatic API inventory sync
  Linking Azure API Management to Azure API Center automatically syncs APIM's APIs into the inventory, updating it without manual effort and enabling free Standard plan features such as linting.
• AsyncAPI Code Generation with Corvus: Typed Producers
  Corvus' new AsyncAPI code generation for .NET creates typed producer classes that validate message payloads against JSON Schema, ensuring contract consistency without runtime errors.
• Changing the engine while the plane is flying: migrating 60,000 apps under live load
  Azure successfully migrated 60,000 Azure Logic Apps Consumption Integration Accounts to Functions v4 under live load using shadow traffic and a progressive rollout with minimal customer impact.
• AsyncAPI Code Generation with Corvus: Typed Consumers
  Corvus generates typed AsyncAPI consumers that automatically validate messages and handle deserialization errors via an explicit IMessageErrorPolicy, ensuring robust handling of malformed data and transport issues without silent message loss.

💡 Internet of Things

• Azure IoT Edge 1.6 LTS is now available
  Azure IoT Edge 1.6 LTS supports .NET 10 until November 14, 2028, providing a clear upgrade path from the earlier 1.5 LTS version that ends support on November 10, 2026, with updated deployment instructions for seamless migration.

🎓 Learning and Certifications

• Exam AZ-802 Study Guide: Administering Windows Server
  The upcoming AZ-802 beta exam study guide will include firsthand insights and performance metrics from the author's actual test experience, offering practical tips for administering Windows Server.
• Microsoft Applied Skills, “Implement a Real-Time Intelligence solution with Microsoft Fabric”
  Microsoft Applied Skills offers free role-aligned assessments for demonstrating expertise in real-time intelligence solutions using Fabric without requiring certification exams.
• SC-500T00 Update: Trainer Materials Available, Labs to Follow
  The SC-500T00 program has released trainer materials and associated Learn content early to give educators more time to prepare, with hands-on labs scheduled for release on August 7.
• Recent and upcoming Microsoft exam changes – Tuesday 21 July 2026
  The article lists recent and upcoming changes to several Microsoft certification exams, including updates scheduled for July 27 and 28, 2026, affecting exams like AZ-400 and SC-100.
• New Microsoft Certified: Multi-Agent AI Solutions Expert Certification
  The certification validates expertise in designing, building, and operating scalable multi-agent AI solutions on Azure using the Agent Framework.
• Microsoft DP Exams: A Data Student’s Almanac
  This almanac provides practical insights for approaching Microsoft DP exams by emphasizing core concepts in technologies such as SQL Server, Azure SQL, and Cosmos DB to help data students navigate diverse data technologies effectively.

⚖️ Management and Governance

• Azure Environment Design: Stop Paying Production Prices for Dev
  Azure Environment Design advises separating non-production environments from production to use cheaper pricing tiers like Azure Dev/Test, emphasizing that identical architecture doesn’t require matching SKUs in dev, test, QA, and pre-prod.
• Azure Service Health has four alert types. You probably configured one.
  Azure Service Health provides four key alert types-Service Issues, Health Advisories, Planned Maintenance, and Security Advisories-that startups frequently overlook, with the latter three essential for preemptively addressing operational risks.
• Understanding billing for the Azure Copilot Observability Agent
  The Azure Copilot Observability Agent operates on a consumption-based model using Azure Agent Credits, capping each operation at 500 AACs, billed per resource or agent subscription.
• Hypervelocity Engineering: Accelerating Enterprise AI with Azure AI Landing Zones
  Hypervelocity Engineering speeds up enterprise AI by combining Azure AI Landing Zones with an AI-driven platform engineering model for continuous governance, fast experimentation, and scalable deployment.
• Azure Lighthouse: Bring Your Partner In, Without Letting Their Identities In (part 2 of 3)
  Azure Lighthouse introduces a secure way for partners to manage Azure resources without granting them full tenant access or identity governance responsibilities.
• Take command of your Microsoft Foundry AI agents with the Azure Copilot Observability Agent
  The Azure Copilot Observability Agent monitors and investigates Microsoft Foundry AI agents' failures, latency, token usage, and backend performance with integrated analysis for root cause identification.
• Should You Use the New Microsoft Entra Tenant Governance or Azure Lighthouse? (part 3 of 3)
  Microsoft Entra Tenant Governance enables secure administration of another tenant's directory and resources, enhancing Azure Lighthouse with identity-based access for tasks such as managing identities and conducting security operations.
• Your startup doesn't have an SRE team. Now what?
  Azure SRE Agent offers custom skills like Well-Architected Framework reviews, compliance audits, capacity planning, FinOps analysis, postmortem generation, Defender for Cloud monitoring, governance assessments, and AI OpenAI posture checks to help startups proactively manage risk without impacting engineering bandwidth.
• Public Preview: Azure DDoS Protection custom policy
  Azure DDoS Protection custom policy in public preview lets customers set tailored mitigation thresholds for individual load balancer IPs, enhancing security granularity and response efficiency.
• Upgrade from SCOM 2022 UR3 to SCOM 2025 Fails Due to Unix MP Compatibility Issue
  The upgrade failure results from missing module definitions and a regression in the ValidityYears parameter between SCOM 2022 UR3 and SCOM 2025 RTM, causing compatibility issues with the Microsoft.Unix.Library management pack.

🚌 Migration

• Moving Petabytes Without the Panic: At-Scale Storage Assessments and Migrations to Azure
  Azure's new tooling simplifies petabyte data movement to Azure using managed services like Azure Migrate, Storage Mover, and Data Box, reducing complexity and downtime.
• SQL Server 2016 end of support is here: Plan your next steps
  SQL Server 2016's end of support requires organizations to plan for reduced security and assistance, considering migration to Azure SQL, upgrade to SQL Server 2025, or deployment of Extended Security Updates (ESU).
• SQL Server 2016 Reaches End of Support: A Customer Engineer's Perspective on What's Next
  SQL Server 2016 ends support on July 14, 2026, necessitating upgrades, migration to Azure SQL Managed Instance, or purchase of Extended Security Updates for continued protection.
• Automating Oracle to SQL Migration Using SSMA Console Skill in GitHub Copilot
  The SSMA Console Skill for GitHub Copilot automates Oracle-to-SQL migrations, reducing manual effort and inconsistency by translating natural language prompts into orchestrated migration steps using SQL Server Migration Assistant (SSMA), enabling teams to industrialize migrations and accelerate outcomes across large portfolios.

🌐 Networking

• Introducing Azure Front Door edge actions - Bringing secure, programmable logic to the edge
  Azure Front Door now provides edge actions enabling developers to run lightweight JavaScript functions globally for real-time personalization, security, and resiliency without added backend complexity.
• Public Preview: Azure Front Door edge actions
  Azure Front Door’s new edge actions enable serverless JavaScript functions for making real-time routing decisions directly at the edge, reducing latency for global users.
• Network Security Perimeter for Azure Event Hubs: Hardening Your Data Streams
  Network Security Perimeter for Azure Event Hubs offers a declarative, organization-wide method to isolate and manage access to event streaming data, improving compliance and efficiency through centralized security policies.
• Generally Available: IPv6 support for Azure VPN Gateway
  Azure VPN Gateway now provides general availability of IPv6 support for dual-stack configurations and seamless IPv6 traffic routing through VPN tunnels using any production SKU gateway with a standard public IP.
• Design the Network Before You Deploy: Best Practices for Microsoft Foundry Standard Agents BYOVNet
  Designing the network early for Microsoft Foundry Standard agents in a Bring-Your-Own VNet is crucial to avoid costly redeployments; proper planning covering subnets, CIDRs, DNS zones, and firewall rules ensures smooth deployments.
• Public Preview: Standard service endpoint
  The new Standard service endpoint in public preview enables seamless, secure connections from IaaS workloads to Azure PaaS services via Private Link, addressing scalability and management challenges.
• Announcing public preview of Azure DDoS Protection custom policy
  Azure DDoS Protection now offers a public preview custom policy for fine-tuning mitigation based on protocols and resources, aligning protection more closely with specific application traffic patterns and future events.

🔐 Security

• Building toward an Agentic SOC: A Portable, Autonomous Malware Investigation Agent
  The portable malware investigation agent streamlines SOC workflows by automating analysis from detection to verdict, enhancing efficiency and tool integration.
• Defending SaaS-based applications against ShinyHunters OAuth abuse
  Threat actors exploited OAuth trust relationships using ShinyHunters tactics to gain unauthorized access, exfiltrate Salesforce data, and maintain persistence across multiple industries.
• Stay in Control of your Data Customer Lockbox for Microsoft Azure
  Customer Lockbox restricts Microsoft engineers to accessing sensitive Azure data only with explicit customer approval in rare critical support scenarios, ensuring essential governance for regulated industries.
• The 3 Azure Security Mistakes I See During Audits
  The article highlights three key Azure security failures: poor monitoring, inadequate alerting, and insufficient response plans that many organizations overlook.
• Hunting Local AI Tools on macOS with Microsoft Defender for Endpoint
  The article explains how to use Microsoft Defender for Endpoint’s new capabilities to inventory and detect local AI tool usage on macOS, addressing security challenges posed by developers running models locally.
• ASP.NET, Kestrel and Schannel GA with TLS 1.3 Post-Quantum Cryptography
  Windows now generally available with hybrid TLS using ML-KEM key exchange groups for TLS 1.3 and post-quantum cryptography, enhancing cryptographic agility against future quantum attacks.
• Announcing Trusted Launch for Virtual Machines for Windows Server Insiders
  Trusted Launch secures Windows Server Insider VMs with Secure Boot, vTPM protection, and boot integrity verification for safe startup and automatic state preservation during failover clusters.
• New Privileged Token Context Telemetry Boosts Advanced Hunting in Microsoft Defender
  Microsoft Defender now highlights key privilege context from token creation events for better threat hunting and detection.
• Announcing the ASIM Parser Creation Agentic Experience
  The new AI-guided agentic experience streamlines ASIM parser creation with structured prompts, automated validation loops, and seamless integration into Azure Sentinel’s open-source contribution process.
• Turning threat intelligence into decisive action with Defender Experts
  Microsoft Defender Experts Threat Intelligence provides tailored expert guidance to help security teams prioritize and respond to emerging threats effectively.
• Microsoft's Record July Patch Tuesday Fixes 570 Flaws, Including Two Exploited Zero-Days
  This month’s Patch Tuesday addressed 570 vulnerabilities, with two zero-days already exploited, highlighting heightened security urgency.
• Microsoft at Black Hat USA 2026: Defending trust in the age of AI and supply chain attacks
  At Black Hat USA 2026, Microsoft highlighted how threat actors target trusted software, services, and AI systems to scale attacks, stressing the need for early detection and collaboration in security operations.
• Microsoft Defender CSPM Assessments for Azure Database for PostgreSQL Flexible Server - GA
  The general availability of Microsoft Defender CSPM assessments for Azure Database for PostgreSQL Flexible Server offers continuous evaluation against PostgreSQL security best practices with actionable recommendations to enhance security and compliance.
• How to use the GPO Policy Analyzer
  GPO Policy Analyzer automates gap analysis by normalizing registry.pol files from multiple Group Policy Objects, highlighting conflicts, missing settings, and redundant entries to help administrators quickly identify drift without touching production environments.
• Setting up Microsoft MDASH (Multi-model Agentic Scanning Harness) Part 1
  Microsoft's MDASH achieves 96.55% accuracy on the CyberGym benchmark by integrating multiple AI models and agents to detect vulnerabilities in codebases, providing confidence scoring, AI-generated fixes, and seamless integration with Azure Foundry and GitHub for automated vulnerability detection within CI/CD pipelines.
• Strengthening Key Management Service - (KMS) with Hardware-Based Trust
  Microsoft enhances its Key Management Service (KMS) with TPM attestation for secure device activation.
• Real world incident response: Microsoft and AXA XL strengthen cyber resilience
  Microsoft and AXA XL partner to offer rapid, coordinated cybersecurity incident response by aligning security, executive, legal, and insurance teams for enhanced speed and effective collaboration.
• Some tools and techniques for hardening Windows Server
  This article details role-specific Windows Server 2025 baselines using OSConfig to harden network exposure, credentials, and resistance to lateral movement and persistence attacks.
• Offboarding from Microsoft Defender for Endpoint
  Microsoft Defender for Endpoint maintains security policies on devices after employee departure due to offboarding requiring a specific signed script rather than just disconnecting from the domain or using Intune.
• Email threat landscape: Q2 2026 trends and insights
  The Q2 2026 email threat landscape saw a 92% drop in Tycoon2FA-linked phishing after its disruption, with Teams-based social engineering-particularly voice phishing-growing nearly tenfold.
• Extending Sentinel Data Integration: Azure Blob Storage Support for CCF Connectors
  The new Azure Blob Storage support in CCF connectors enables ISVs to securely deliver security data into Sentinel, improving resilience and scalability for high-volume ingestion.
• Behind the Build with Gigamon: Enriching Microsoft Sentinel with Network-Derived Telemetry
  Gigamon and Microsoft integrated network-derived telemetry into Sentinel to enhance security investigations with deeper context and AI-driven insights for comprehensive threat detection in hybrid environments.
• Operationalizing the Agent Blast Radius Model: Policy Gates, Tool Contracts, and Rollback Controls
  The article details how to apply the Agent Blast Radius Model through policy gates, limited tool contracts, approval processes, rollback protocols, and audit trails for enhanced AI safety.
• How to Keep Secrets and Sensitive Data Out of AI Prompts, Embeddings, Traces, and Logs
  The guide details a framework to prevent sensitive information leakage in AI systems through early classification, blocking credentials, redaction, encryption, and strict retention policies.

📦 Storage

• RELEASE: Microsoft 365 Archive now available in Government Community Cloud (GCC)
  Microsoft 365 Archive now allows GCC tenants to archive inactive SharePoint sites via an admin center, offering a pay-as-you-go model that cuts storage costs while maintaining compliance and searchability.
• When Compute Sits and Waits: Fixing the Hidden Storage Bottleneck in EDA and HPC with Azure NetApp Files
  Azure NetApp Files resolves hidden storage bottlenecks in EDA and HPC through native Azure integration with NetApp ONTAP, high concurrency I/O support, and migration tools such as SnapMirror.
• Azure Elastic SAN: Pooled, Cloud-Native Block Storage That Actually Acts Like a SAN
  Azure Elastic SAN provides a pooled, cloud-native block storage solution that emulates traditional SAN architecture, enabling efficient provisioning of shared capacity and allocation across multiple workloads.
• Premium SSD v2 and Instant Access Snapshots: A Better, Faster, Cheaper Disk for Your Azure VMs
  Premium SSD v2 provides four times the IOPS and twice the throughput of its predecessor at 42% lower cost, featuring live resizing and instant access snapshots to cut restore times and operational overhead.
• Feeding the GPUs: File Storage for AI and Cloud-Native Workloads on Azure
  Azure offers tiered file storage solutions-Managed Lustre for high-throughput training, Azure Files for fast ReadWriteMany inferencing, and Blob Storage as the cost-effective core-to optimize GPU utilization and reduce costs for AI workloads on Azure.
• Accelerating cross-region Azure disk copying
  Azure has optimized cross-region disk copying with AzCopy often completing transfers under a minute faster than snapshots or BLOB copies.
• Maybe I Owe Azure Backup an Apology...
  The article warns against judging Azure Backup too harshly based on outdated technical opinions stemming from past experiences.
• Cut Your Azure Blob Storage Bill in Half: A Practical Walkthrough of Object Storage TCO
  Cut Your Azure Blob Storage Bill in Half offers practical steps to optimize object storage costs by understanding capacity, transactions, retrieval fees, and network egress, reducing expenses without altering applications.
• Azure Files, Reimagined: Top-Level Shares with Per-Share Networking, Billing, and Scale
  Azure Files now provides top-level file shares with independent RBAC, networking, billing, IOPS, and throughput per share, simplifying management and reducing latency.
• Faster blob copying in Azure Blob Storage
  Azure Blob Storage's new StartCopyFromUriAsync feature speeds up server-side copying of large files to complete in minutes, surpassing simple download-and-re-upload methods though remaining slower than parallel block-from-URL approaches.
• No, That Vault Isn't What You Think It Is
  Microsoft's use of "Recovery Services Vault" for both Azure Backup and the less utilized "Backup Vault" has led to considerable confusion among architects and users.

💻 Virtual Desktop Infrastructure

• Modernize VDI with Azure Files and Entra Cloud-Native Identities
  Azure Files now lets VDI users authenticate directly with Microsoft Entra ID over SMB Kerberos, eliminating the need for on-premises domain controllers and enabling cloud-native identity management for persistent profiles.
• Automating the Disconnected Session Time Limit Policy with PowerShell and Microsoft Graph
  Automating the disconnected session time limit policy with PowerShell and Microsoft Graph ensures consistent capacity management across all host pools, saving administrative effort.

🔗 Web

• Hosting Python APIs on Azure App Service — Linux Only, Startup Commands, ODBC Driver Names, and Installing Drivers from SSH
  Azure App Service now supports only Linux for Python APIs, eliminating Windows support and requiring developers to use built-in runtimes, configure startup commands, and specify ODBC drivers in connection strings.
• Connect Azure App Service to Azure SQL Database Using a User-Assigned Managed Identity (Python + pyodbc)
  This article shows how to securely connect an Azure App Service hosting a Python FastAPI app to Azure SQL Database via pyodbc using a User-Assigned Managed Identity, removing password requirements and lowering risk of outages due to credential changes.
• 351 - Building and hosting a static site in 2026
  The episode explains moving from Ghost and WordPress to an Astro-built static site hosted on Azure Static Web Apps, emphasizing lessons gained via AI-assisted development and hosting plan choices.
• Give your AI agent two memories with Azure App Service
  Azure App Service uses Azure Managed Redis and Cosmos DB to enable AI agents with separate short-term conversation histories and long-term user-scoped durable memories, optimizing cost, latency, privacy, and deletion through purpose-built stores integrated via the Agent Framework context pipeline.
• Microsoft Foundry Now Has an AI Gateway Control Plane — What Changes for App Service
  Microsoft Foundry now includes an AI Gateway control plane in its Admin console, allowing platform teams to create or associate APIM instances, set independent token limits per project, and manage traffic governance directly from Foundry, streamlining AI adoption while keeping APIM as the underlying traffic gateway.
• Azure App Service Pitfalls: Why a PostgreSQL Migration Fails in Production
  The article describes a PostgreSQL migration to Azure App Service that moved data successfully but failed in production due to unforeseen platform configuration and networking issues, providing insights on diagnosis and resolution.
• MCP in Azure: Using API Management for Authentication, Access, Logging & Governance
  Azure API Management provides governance, identity management, access control, traffic monitoring, and lifecycle policies for MCP servers, ensuring consistent security and operational standards.
• Proactive Resiliency with Azure App Service Auto Heal
  Azure App Service’s Auto Heal automatically detects and recycles unhealthy worker processes, maintaining availability as per the Azure Well-Architected Framework.
• Beyond text: Returning images and interactive apps from MCP servers
  The Model Context Protocol server shows how to use its tools to search a nature photo library, display thumbnail images in VS Code, and create an interactive gallery, highlighting its multimodal functionality.