Azure Weekly
Issue 588
15th November 2026
.NET 11 was released this week, with Day 1 support on Azure Functions. This milestone also means .NET 8 & 9 go out of support. The Developer Tools section is particularly busy this week with related updates.
- .NET 11: unions, MAUI ready for business, but what comes next? by Tim Anderson - .NET 11 introduces union types, faster async code, Core CLR in MAUI, and performance boosts across runtimes, making it a strong choice for cross-platform mobile and desktop development.
- .NET 11 is now generally available in Azure Functions by Thiago Almeida - Azure Functions now supports .NET 11 in the isolated worker model with a new Azure.Functions.Sdk package simplifying project setup.
- Retirement: Support for .NET 8 and .NET 9 on Static Web Apps is ending November 10, 2026 by The Azure Updates Team - Static Web Apps will cease supporting .NET 8 and .NET 9 on November 10, 2026, necessitating migration to .NET 10 for future security updates and support.
Other highlights include:
- What’s new with Microsoft in open-source and Kubernetes at KubeCon North America 2026 by Brendan Burns - At KubeCon + CloudNativeCon North America 2026, Microsoft announced improvements to Kubernetes including GPU handling via device plugins, real-time cluster drift monitoring across fleets, and an agent infrastructure for secure portable execution on multiple clouds.
- Claude Fable 5.2 and Mythos 5.2 bring frontier intelligence to Microsoft Foundry by Amar Badal - Claude Fable 5.2 and Mythos 5.2 enhance Microsoft Foundry with advanced capabilities for complex tasks in software engineering, scientific research, and finance, featuring longer workflows and specialized support for cybersecurity and life sciences.
- Generally Available: Azure Front Door edge actions by The Azure Updates Team - Azure Front Door now offers generally available edge actions, enabling serverless execution of lightweight JavaScript functions directly at the edge for real-time request and response customization. Edge actions are built on Hyperlight, Microsoft's secure micro-VM technology, so each action runs in its own hardware-isolated micro-VM. For more background, see Introducing Azure Front Door edge actions - Bringing secure, programmable logic to the edge from the preview launch in July.
And a reminder that Microsoft Ignite runs from 17th to 20th November. The Learning and Certifications section has session guides for PostgreSQL, SQL, Arc, Defender, Sentinel and more.
🤖 AI
- Notes of the week: managed agent harnesses, Microsoft's own models and Claude in the EU The managed Copilot harness in Foundry now runs the agent loop and sandbox, offering reusable skills and integration with Azure tools, while expanding model options to include both Microsoft AI's frontier models and Anthropic’s Claude models within the EU Data Boundary for enhanced compliance.
- From Python Agents to Copilot Harness : Putting Microsoft Foundry to the Test The experiment demonstrates moving a Python/MAF research demo into a Foundry-managed Copilot harness, preserving authentication, streaming, and secure access with minimal custom code.
- Real-Time Multichannel Speech-to-Text is now generally available Azure Speech's real-time multichannel speech-to-text now includes language identification and channel separation for efficient transcription of mixed-language conversations without additional costs.
- Claude Fable 5.2 and Mythos 5.2 bring frontier intelligence to Microsoft Foundry Claude Fable 5.2 and Mythos 5.2 enhance Microsoft Foundry with advanced capabilities for complex tasks in software engineering, scientific research, and finance, featuring longer workflows and specialized support for cybersecurity and life sciences.
- Foundry Toolbox Lifecycle: Deploy and Promote Managing Foundry Toolboxes in separate azure.yaml files allows efficient reuse of integrations with centralized tool governance via a single MCP-compatible endpoint.
- Native real-time multilingual transcription is now generally available, and faster Azure Speech now provides generally available native real-time multilingual transcription with results appearing in about 100 milliseconds across 61 locales using an LLM speech streaming model.
- New Microsoft Copilot Harness | More power, token efficiency & control for agents in Foundry The new Microsoft Copilot harness converts Foundry prompt agents into autonomous planners that loop, fetch data, and produce executive reports with enterprise controls such as unique Entra Agent IDs, egress limits, and enhanced token efficiency via tool search.
- Azure Human Approvals: Stop the Agent Before the Tool Fires Azure Human Approvals adds a built-in gate within the Foundry Agent Service, allowing automated workflows to pause execution and require Entera ID-verified approval before proceeding.
- Azure Content Understanding Azure Content Understanding integrates OCR and a generative LLM into one call to extract typed values with confidence scores and page locations from documents, simplifying data extraction workflows.
- Introducing Kimi K3.1 Pro through Fireworks on Microsoft Foundry Kimi K3.1 Pro with 2.8 trillion parameters and a 1-million-token context window is now accessible via Fireworks on Azure’s Foundry for use in software development, research analysis, and 3D design tasks.
- Build a dual-mode voice agent in Microsoft Foundry with native subagents Microsoft Foundry uses a dual-mode pattern to separate real-time conversational flow from complex backend tasks, ensuring agents remain responsive during specialized processing.
- Multilingual Fabric data agents in Foundry: normalize values Normalizing Chinese values in Foundry data agents boosts accuracy from 28% to nearly 70%, fixing issues with Traditional script mismatches against Simplified records.
- Best practices for building agentic AI applications with Foundry Agent Service Foundry Agent Service simplifies building reliable, secure, and cost-effective agentic AI applications by emphasizing core engineering principles like starting with the business problem and defining success early.
- Available today: Anthropic Claude Fable 5.2 in Microsoft Copilot Claude Fable 5.2 in Microsoft Copilot integrates Work IQ to ground responses in your organization's files, meetings, and data, enhancing accuracy and trustworthiness for complex tasks.
- Where do I govern my Copilot Studio agents? The blog details five key governance outcomes to help IT teams manage Copilot Studio agents securely and efficiently.
- Sovereign AI, From Azure to the Edge: My NVIDIA GTC Berlin 2026 Session Is Now Available The recording covers how sovereign AI enables secure, localized data processing from Azure cloud to edge devices, highlighting practical use cases and integration strategies.
- Copilot Studio | The Copy It Calls (Upgraded) Copilot Studio's new upgrade allows users to naturally request an agent transition to the GitHub Copilot harness, retaining instructions and tools while excluding knowledge sources, Power Fx expressions, and child agents.
- Device economics: Making Surface compute count Surface devices help organizations balance AI workloads between endpoints and the cloud, optimizing costs by aligning model complexity with device capabilities like CPUs, GPUs, NPUs, and memory.
- NVIDIA BioNeMo Agent Toolkit: Bringing NVIDIA Models in Biology to Every Scientist The NVIDIA BioNeMo Agent Toolkit allows scientists to design protein binders through a natural language interface on Azure's Discovery platform, utilizing the RFD4-Proteina AI model for simultaneous generation of 3D structures and sequences, thereby speeding up drug discovery processes.
🔎 Analytics
- Retirement: Azure Data Share will be retired on November 12, 2029 Azure Data Share will retire on November 12, 2029, encouraging users to move to Azure's new Microsoft Fabric platform for improved data sharing capabilities.
- From Files to Insights: Building a Medallion Architecture with Microsoft Fabric Microsoft Fabric's medallion architecture converts raw vendor files in a SharePoint folder into clean, trusted data for analytics via Dataflows that ingest, filter, and map Excel content into a unified Lakehouse storage solution.
- Automate Databricks Unity Catalog Access via Terraform Automate Databricks Unity Catalog access using Terraform’s entitlement matrix, import existing permissions, and enforce plan gates to prevent unauthorized access changes.
- Conditional Retries in Microsoft Fabric Pipelines Implement conditional retries in Azure Data Factory pipelines to automatically handle specific activity errors, improving workflow reliability without manual intervention.
- Using AI to Build Semantic Models, Ontologies, and Fabric Data Agents AI builds semantic models, ontologies, and Fabric Data Agents in Microsoft Fabric to clarify data meaning and enhance natural language analytics.
- Server-side Fabric App functions Azure IoT Operations introduces Server-side Fabric App functions, enabling seamless command execution via Event Grid Namespace and secure connection secrets management within Fabric Apps.
- A next step in the evolution of the Microsoft Data Governance strategy The next chapter of Microsoft’s data governance strategy integrates OneLake catalog into Azure Fabric to keep policies, trust signals, and lineage near the data analyzed by AI agents.
- From Purview Data Governance to the Fabric OneLake Catalog The transition from Purview Data Governance to Fabric OneLake Catalog in Microsoft Fabric redefines data governance architecture, with OneLake serving as the primary governance experience while Purview handles security and compliance, offering a clear rule of thumb for managing governance responsibilities.
🖥️ Compute
- Generally Available: Azure Functions support for .NET 11 Azure Functions now provides general availability for .NET 11 with an isolated worker model for building serverless applications.
- Your experience matters: help shape what’s next for Windows Server The Windows Server TAP lets organizations test preview features in non-production settings to help shape future enterprise improvements.
- Accelerating AI Data Processing, HPC, and Agentic Workloads with Azure HDv2 Virtual Machines Azure HDv2 VMs powered by 6th Gen AMD EPYC processors offer unprecedented speed and energy efficiency for AI data processing with over 500 CPU cores, 3.8 TB DDR5 memory, and 30 TB PCIe NVMe storage.
- .NET 11 is now generally available in Azure Functions Azure Functions now supports .NET 11 in the isolated worker model with a new Azure.Functions.Sdk package simplifying project setup.
- Azure VMs deliver inference performance in line with industry benchmarks Azure ND GB200 v6 VMs perform well in inference benchmarks with higher throughput and faster first tokens while matching decode and end-to-end latency of bare metal servers.
- From Silicon to Service: Bringing AMD Helios™ to Azure Azure’s new ND MI455X v7 AI supercomputing cluster integrates AMD Helios hardware with a two-tier Ethernet network to deliver high-availability, low-latency inference for demanding AI workloads.
🚢 Containers
- AKS Upgrade Insights Now in Preview The AKS Upgrade Insights API in preview offers a dry run feature for Kubernetes upgrades, allowing teams to test changes without affecting production workloads.
- Talos on Proxmox: What Actually Worked (and the Boot Loop That Taught Me Why) Building a highly available Kubernetes lab on Proxmox revealed successful configurations and challenges such as boot loops and network drops, ultimately resulting in a robust cluster and useful automation scripts.
- Azure Container Linux (ACL): making immutable infrastructure practical at scale Azure Container Linux introduces simpler fleet scaling, hardware-backed protection, and enhanced code integrity tools for scalable immutable infrastructure in Azure Kubernetes Service.
- What’s new with Microsoft in open-source and Kubernetes at KubeCon North America 2026 At KubeCon + CloudNativeCon North America 2026, Microsoft announced improvements to Kubernetes including GPU handling via device plugins, real-time cluster drift monitoring across fleets, and an agent infrastructure for secure portable execution on multiple clouds.
- What’s new in AKS networking: Connectivity, performance, and observability Azure Kubernetes Service adds cross-cluster networking, Multi-NIC pods for AI workloads, and Managed DRANET support for RDMA NICs to simplify multi-cluster connectivity and improve performance.
- 368 - Azure Container Apps Sandboxes: a microVM for every agent task Azure Container Apps Sandboxes offer enhanced security and isolation through microVM environments for each agent task, integrating seamlessly with Azure’s networking and egress policies.
- Azure Day with Kubernetes at KubeCon NA 2026: My AKS Highlights The event focused on Kubernetes with Azure Kubernetes Service (AKS), showcasing new capabilities and integrations.
- Deploying an AI Application on Azure Container Apps Deploying an AI app on Azure Container Apps involves packaging, storing images securely in Azure Container Registry, integrating with Azure OpenAI for inference, and automating scaling through Azure Container Apps Environment.
- Container Strategy: Running Containers and Actually Governing Them Are Two Different Things Kubernetes sprawl creates significant security and compliance gaps that require centralized policy enforcement beyond simple container deployment.
🗄️ Databases
- Public Preview: SQL performance monitoring for Azure SQL Managed Instance Azure SQL Managed Instance’s new public preview offers built-in SQL performance monitoring, reducing the need for custom scripts and disconnected tools for businesses with critical workloads.
- Table Partitioning in SQL Server: A Practical Guide to Managing Very Large Tables (VLDBs) SQL Server’s table partitioning lets you split large tables into manageable parts based on columns like order date, improving maintenance and performance as data volumes grow to billions of rows.
- An Index Scan Is Only Half the Story: On-Disk Ordering Affects Performance On-disk ordering can dramatically affect performance of an Index Scan, even when the index itself appears healthy, by increasing the number of heap pages PostgreSQL must visit to retrieve rows.
- Closing a monitoring blind spot: version store read metrics Azure SQL Database now provides version store read metrics to detect performance issues caused by prolonged row versions retention, helping diagnose CPU spikes without needing XPerf traces.
- Restore Database Mistake: I Learned About Databases From That This account describes a restore database error caused by using REPLACE without checks, resulting in the accidental restoration of production instead of test, emphasizing the importance of verifying server names and avoiding REPLACE on live systems.
- Private Database, Public Tools - Securing Azure Database for PostgreSQL Flexible Server Behind MCP Securing Azure Database for PostgreSQL Flexible Server behind Managed Connectivity Platform involves addressing authentication, authorization, and operational controls alongside private network connectivity to prevent unauthorized access.
- Enforcing Unique Customer Emails When Duplicates Already Exist SQL Server filtered unique indexes allow multiple NULL values while enforcing uniqueness on existing emails after you archive duplicates using ROW_NUMBER to identify the most recent records.
- How SQL Server People Can Stay Relevant With AI Erik Darling advises SQL Server professionals to focus on practical applications such as performance monitoring and robot-assisted data analysis for staying relevant in the AI era instead of creating low-quality content with AI tools.
- PostgreSQL 18 EXPLAIN Improvements PostgreSQL 18 improves EXPLAIN output by displaying buffer activity, index search counts, and detailed memory/disk usage to better identify query performance issues.
- Ask your database estate a question: Copilot in Database Hub is now in public preview Copilot in Database Hub enables database teams to ask natural-language questions about performance, security, and migration readiness, offering instant summaries and detailed insights across Azure, Fabric, on-premises, and multicloud databases.
- SQL Server Performance Consulting for the AI Era Erik Darling's workshop teaches developers and DBAs to collaborate with AI agents for SQL Server performance tuning while addressing accuracy and knowledge integration challenges.
- PostgreSQL Bloated Tables PostgreSQL table bloat happens when MVCC leaves dead rows unremoved, causing wasted space and slower queries until cleaned by VACUUM or tools like pg_repack.
- Build faster apps: database caching patterns with Azure Managed Redis Azure Managed Redis helps fast apps by caching hot data from databases like PostgreSQL, SQL Database, and Cosmos DB, reducing latency and handling traffic spikes with sub-millisecond reads.
- Microsoft.Data.SqlClient 7.1.2, 7.0.4, and 6.1.8 are now available The new SqlClient servicing releases enhance security by requiring HostNameInCertificate values to match server certificates, preventing TLS handshake failures due to misconfigurations.
- EF 11 Drops a Redundant Join From Split Queries EF 11 removes unnecessary joins from split queries, reducing database load and improving performance by up to 24 percent in PostgreSQL benchmarks with four reference navigations included.
- Lessons Learned #563: How to Check Which Backups Are Available in Azure SQL Database This guide explains how to verify Azure SQL Database backup availability by checking service tier, retention settings, and using tools like the Azure portal or CLI queries on sys.dm_database_backups for detailed metadata.
🛠️ Developer tools
- Durable Task extension for Microsoft Agent Framework is now Generally Available The Durable Task extension for the Microsoft Agent Framework is now generally available, enabling automatic recovery from failures while preserving conversation state and workflow progress without manual intervention.
- GitHub Maps AI Agents Across the Full Software Development Cycle GitHub Copilot is now integrated into every stage of the software development lifecycle, from planning to production troubleshooting, according to a new comprehensive guide.
- Copilot in Visual Studio Just Got a Major Upgrade Visual Studio’s Copilot now integrates GitHub’s Copilot SDK, improving performance by handling 21 extra tasks in .NET projects and providing more precise, controllable answers with live credit monitoring.
- MCP server vs REST API: When should you use each? MCP servers suit AI agents requiring standardized tool discovery, whereas REST APIs perform best in traditional applications with predictable client-server interactions and robust infrastructure.
- Where Should the LLM Sit? Lessons from a Local Clinical Ranking Prototype with Foundry Local The study demonstrates that incorporating deterministic normalization and a disease-symptom graph before local LLM reranking enhances diagnostic ranking accuracy to 64.8% Hit@1 compared to direct model calls.
- Announcing Visual Studio 2026 LTSC 18.11 Visual Studio 2026 LTSC 18.11 now offers two years of support, extending through November 14, 2028, to give organizations more time for validation and deployment after the initial one-year period.
- .NET 11 Ships with C# 15 Unions, Faster Tools and CoreCLR for MAUI .NET 11 adds C# 15 unions for compiler-checked models, boosting tool performance and optimizing CoreCLR for MAUI apps.
- .NET 11 is here. Should you upgrade, and what does it bring to AI? .NET 11 adds C# 15 features such as union types and collection expression arguments, along with faster builds via Native AOT and enhanced test automation, plus new Blazor AI components for interactive service desk agents.
- Engineering an Agentic Scheduler Engineering an Agentic Scheduler explains how Microsoft created a system with LLMs that handle complex scheduling constraints deterministically through stepwise processes and structured inputs.
- Running GitHub Copilot, Claude Code and Codex in VS Code dev containers VS Code dev containers allow developers to define custom environments for GitHub Copilot, Claude Code, and Codex, controlling installed tools and file access directly within the repository.
- .NET 11: unions, MAUI ready for business, but what comes next? .NET 11 introduces union types, faster async code, Core CLR in MAUI, and performance boosts across runtimes, making it a strong choice for cross-platform mobile and desktop development.
- Build a Chat App with the Microsoft Foundry SDK: From Playground to Python for AI-901 The video shows how to convert a gpt-5-mini deployment in the Foundry playground into a working Python chat app with the Microsoft Foundry SDK, adding conversation memory, real-time streaming responses, and keyless sign-in through Azure Entra ID.
- Python support has arrived in the guided Copilot experience for VS Code The guided Copilot experience for VS Code now fully supports Python, allowing developers to plan, scaffold, develop locally, and deploy Azure apps with a structured workflow tailored for Python projects.
- Bring Your Own Model is now generally available in Visual Studio Bring Your Own Model in Visual Studio 18.12 enables developers to utilize private, API-keyed, or locally hosted AI models within the IDE, supporting multiple provider configurations and custom endpoints with enhanced proxy handling for enterprise networks and an intuitive setup process.
- Uno Platform Studio 3.3: Your AI Workbench, Now on Desktop and Mobile Uno Platform Studio 3.3 adds a mobile companion app for testing apps on real devices with OTA updates, boosting developer-AI collaboration in the .NET ecosystem.
- What I think so far about Spec Driven Development Spec Driven Development combines detailed specifications with AI to generate code and artifacts, raising questions about the model's role as source of truth versus evolving alongside it.
- Building a Privacy-Aware Personal Agent with Microsoft Foundry Local Building a Privacy-Aware Personal Agent with Microsoft Foundry Local shows how to keep sensitive data on-device while optionally using the cloud for tasks, preserving citations and enforcing strict privacy boundaries via sanitization and deterministic retrieval.
- Uno Platform 7.0: 2x faster, 5x smaller, built for the agentic era Uno Platform 7.0 offers about twice the speed and five times less size than its predecessor while maintaining seamless compatibility with .NET MAUI controls across all platforms including Apple TV and Linux thanks to a unified rendering pipeline powered by SkiaSharp 4 and an optional WebGPU backend.
- Visual Studio 18.12 Makes Bring Your Own Model Generally Available Visual Studio 18.12 introduces Bring Your Own Model support, allowing developers to seamlessly integrate local, hosted, and enterprise AI models using enhanced configuration settings within the IDE.
- Build once, run on any agent: A practical guide to agent skills Agent skills allow teams to capture and reuse repeatable instructions via a versioned directory centered on a SKILL.md file that describes tasks, separating tool access from instruction sharing.
- VS Code 1.146 Adds AI Agent Inbox, Model Comparisons and Chat Windows VS Code 1.146 adds an AI Agent Inbox, model comparison tools, customizable chat windows, and improved security while retiring the Local agent harness in February.
- Introducing Azure Container Apps Sandboxes in Visual Studio Code The new Azure Container Apps extension for Visual Studio Code integrates sandbox groups and individual sandboxes into the Azure Resources view, enabling developers to create, manage, snapshot, stop, and resume containers directly from their IDE.
- What’s New in .NET 11: Features, C# 15 and Practical Examples .NET 11 introduces C# 15's union types and other language enhancements, including closed hierarchies and collection expression arguments, with two years of support versus one year for .NET 10.
🔩 DevOps
- GitHub Advanced Security Now Has a Self-Serve 30 Day Trial: What's Free, What Isn't, and How I'd Spend It GitHub Advanced Security’s new self-serve trial offers full access to its threat detection and secret scanning features for 30 days, with limited usage on advanced anomaly detection tools.
- Migrate a GitHub repository to GitHub Enterprise Migrating a GitHub repository to GitHub Enterprise Cloud requires installing tools, creating scoped personal access tokens, configuring SSO, running the migration command via PowerShell, validating the new repo, and deleting the tokens.
- Eliminate Azure DevOps Secrets with Workload Identity Azure DevOps removes client secrets and personal access tokens by integrating workload identity federation, ensuring stolen credentials are detected before going live.
- We couldn't see what we gave back. So we built giveback. Giveback is an open-sourced GitHub App that tracks and reports on contributions made by Michelin engineers to external open source projects using their personal accounts.
- GitHub availability report: October 2026 GitHub doubled Azure Central US compute resources improved reliability reduced outages and database loads but still faces network disruptions and unsafe routing issues.
- Which build is deployed in each Azure DevOps environment? A successful build or an approval wait doesn't tell you what is actually running in each environment; this post walks through the native Azure DevOps Environments views for answering that, and where the free worre.dev Windows app helps when tracking several applications and environments (Azure DevOps Services only, not classic releases or Azure DevOps Server).
- I Deployed Infrastructure to a Commodore 64 Bicep's local deploy feature enables declarative, idempotent infrastructure management on a Commodore 64, allowing code to control hardware registers and run custom programs.
🧬 Hybrid + multicloud
- Azure Local 2610 Azure Local 2610 features rack-aware storage clusters, secure confidential VMs with AMD SEV-SNP, unified OS and SDN patches, and a flexible security configuration framework.
- Azure Local security: closing the industry standards gap and improving enterprise adoption Azure Local version 2610 adds a single security setting to better control drift, aligning with CIS Benchmarks and DISA STIG to enhance enterprise adoption.
🎭 Identity
- What’s New in Entra Private Network Connector v1.5.5130.0 The new Entra Private Network Connector adds a Performance tab with CPU and memory metrics for up to two days, improved DNS per network interface, and enhancements to logging, reliability, and security.
- GSAClientDiagnosticsToolkit is here: a PowerShell module to find out why Global Secure Access isn't working on your device The GSAClientDiagnosticsToolkit PowerShell module helps IT professionals quickly diagnose and resolve issues with Azure's Global Secure Access implementation on individual devices.
- Check out the new Unified Roles preview for Entra ID Microsoft Entra Unified Roles introduces cross-service access and constrained delegation, with the portal code showing planned PIM integration and preview access controls.
- Azure Managed Identity vs Service Principal: Which to Use Compares managed identities and service principals for real workloads, covering the default Contributor role from az webapp identity assign, GitHub OIDC federated credentials, and why role changes can take up to 24 hours to show up in tokens.
🔌 Integration
- From Inbox to Audit Trail: Automating Compliant Email Export with Purview, Sentinel, and Logic Apps The solution automates compliant email export from Exchange Online to SharePoint using Purview for classification, Sentinel for detection, and Logic Apps for orchestration, eliminating manual processes and ensuring auditability.
- agentgateway: an open source AI gateway on Kubernetes agentgateway is an open source AI gateway for Kubernetes that proxies calls to language models and MCP tools, applying rate limits and logging while integrating with GitOps workflows via Argo CD.
- Automate Contract Renewal Reminders Before They Auto-Renew Power Automate creates automated reminder workflows that notify contract owners with escalating alerts-90, 30, and 7 days prior to auto-renewal-to ensure timely action.
- Migrate NuGet-based Logic Apps Standard apps and custom built-in connectors to .NET 10 Migrating NuGet-based Logic Apps Standard apps and custom built-in connectors to .NET 10 requires retargeting projects, resolving dependency conflicts, and adjusting Azure app settings with minimal code changes.
🎓 Learning and Certifications
- Azure Arc at Microsoft Ignite 2026: Sessions, hands-on labs, and ways to connect At this year’s Microsoft Ignite 2026 in San Francisco or online, discover hands-on labs and sessions focused on unifying operations across distributed environments with Azure Arc.
- AI-901 Computer Vision and Image-Generation Models: Analyze and Create Visual Content The AI-901 exam distinguishes between computer vision tasks such as image classification and object detection, which analyze existing images, and image-generation models that create visuals from text prompts, highlighting multimodal capabilities for conversational contexts.
- Microsoft Ignite 2026: Your Guide to Agent 365 Microsoft Ignite 2026's Agent 365 guide showcases the platform's evolution into an AI observability solution for enterprise agent governance, security, and financial visibility.
- Decisions, decisions: Your guide to PostgreSQL at Microsoft Ignite 2026 At this year's Microsoft Ignite 2026 in San Francisco and online from November 17 to 20, PostgreSQL sessions cover modern app development, AI integration, open-source data updates on Azure, service comparisons for workload architecture decisions, and hands-on workshops to help technology leaders make informed choices.
- Microsoft Ignite 2026, lots of IoT, RTI, and IQ news The Ignite 2026 event includes IoT, RTI sessions, a keynote on AI transformation, security of edge AI workloads, and demos of Foundry IQ’s knowledge graph.
- Catch Microsoft Defender and Sentinel at Ignite 2026 At Ignite 2026 in San Francisco, catch key Defender and Sentinel sessions outlining a new agentic SOC blueprint, autonomous protection strategies, and identity risk signal integration.
- Microsoft Sentinel at Ignite 2026: what to see and where to start At Ignite 2026, discover how the new Agentic SOC blueprint integrates Sentinel with Defender and expanded SIEM capabilities for enhanced detection and AI-driven workflows.
- Important Update to the MSLE Lab Benefit The MSLE Lab Benefit is being discontinued for new or pending class requests, so educators cannot request new funded labs, but existing funded classes will receive transition guidance and continue with a discounted alternative.
- Microsoft Ignite 2026: Find Me in San Francisco At Microsoft Ignite 2026 in San Francisco, catch the author's Lightning Talk on applying Ignite insights to real-world impact and join Table Talks on agent governance and RAG transformation.
- Join Intel and Microsoft at Ignite 2026 At Ignite 2026, Intel and Microsoft will unveil innovations such as advanced manufacturing insights, agentic AI, sovereign cloud with confidential computing, and edge AI integrated with robotics.
- Plan your Ignite agenda: Join sessions that help you secure agentic insiders at scale. Discover how Project Perception provides real-time risk protection and dynamic access controls for agentic insiders at Ignite's hands-on labs from November 17-20 in San Francisco and online.
- Join NVIDIA and Microsoft at Ignite 2026 NVIDIA and Microsoft will collaborate across keynote presentations, breakout sessions, hands-on labs, and a dedicated show floor exhibit at this year’s Ignite 2026 event in San Francisco and online.
- Visit Azure and AMD at Microsoft Ignite and SC26 Azure and AMD are collaborating to expand AI infrastructure with purpose-built VMs, managed MI300X endpoints, and upcoming MI455X accelerators.
- Are Certifications Relevant? A Photograph of a River Certifications are largely irrelevant today because AI tools can answer exams, proving only a one-time test pass without showing practical skills or adaptability.
- Join me at the Microsoft Digital Sovereignty Day in Lausanne 2026 The Digital Sovereignty Day in Lausanne 2026 will focus on sovereign private cloud solutions with Azure Local, following successful events in Bern and Zurich.
- Your day-by-day guide to Microsoft SQL at Ignite 2026 At Ignite 2026, Microsoft provides a day-by-day guide on SQL Server modernization, AI integration, architecture choices, and expert-led hands-on development sessions.
- Secure Data at Microsoft Ignite 2026: Plan Your Experience At Microsoft Ignite 2026, discover how Microsoft Purview sessions guide you through data security best practices as you scale Copilot and AI adoption.
- New Course Coming Soon: AI-103 Implementing Text Analysis Solutions AI-103 prepares developers for the Azure AI Apps and Agents Developer Associate exam by covering practical text analysis decisions including API vs. model choices, multilingual data handling, and personal information compliance.
- Sovereign Cloud at Microsoft Ignite 2026: Sessions, Booths and Where to Find Me At Microsoft Ignite 2026, expect in-depth sessions and dedicated booths showcasing Sovereign Private Cloud, sovereign AI with Foundry Local, and Azure Local services focused on digital sovereignty.
⚖️ Management and Governance
- Public Preview: Auto scaling of ingestion limits in Azure Monitor workspace Azure Monitor’s public preview feature lets workspaces automatically adjust ingestion limits up to 10× current volume, optimizing performance during peak usage without manual intervention.
- Combining Cloud computing and Retro Computing! The project automates provisioning of a retro Commodore 64 system using Bicep IaC for seamless integration with modern cloud services.
- Monitor Every GenAI Token with Grafana and Azure Monitor Grafana visualizes every generative AI token consumed, integrating with Azure Monitor via OpenTelemetry to track model and tool usage for precise billing insights.
- Granular control with scoped enforcement actions in Azure Policy for AKS Azure Policy for AKS enables applying enforcement actions like deny, audit, or warn at specific points such as the Kubernetes API server or Gatekeeper webhook for fine-grained policy control.
- Introducing Azure Monitor dashboards with Grafana MCP: Telemetry for AI Agents, Signed In as You Azure Monitor dashboards now integrate Grafana's remote MCP server allowing AI agents such as GitHub Copilot to query Azure resources like Log Analytics, Resource Graph, Prometheus in workspaces, and Azure Data Explorer without extra infrastructure or data sources.
- What's new in FinOps toolkit 15 – October 2026 FinOps toolkit 15 adds an agent plugin for integrating AI assistants with cost data analysis tools, providing real-time FinOps guidance and automating reservation and savings plan management in Azure.
- AI Is Changing Terraform Work—and the Role of Platform Engineers and SREs GitHub Copilot and Claude Code are transforming Terraform workflows on Azure, offering platform engineers and SREs an opportunity to simplify reusing, validating, and operating proven infrastructure patterns.
- What’s new in the Azure Resource Manager MCP server at Ignite 2026 Azure Resource Manager's MCP server at Ignite 2026 introduces expanded write operations, action capabilities, enhanced deployment workflows, single-resource reads, cost tools, and an experimental SKILLS feature for comprehensive infrastructure management.
- Azure Budgets Don't Stop Anything: The Enforcement Gap and How to Close It Why Azure budgets are alerts rather than cost controls, and how to actually stop spend when a limit is reached.
🚌 Migration
- Introducing Azure Container Apps as a target workload in Azure Migrate Azure Container Apps becomes a target workload in Azure Migrate allowing organizations to assess web apps for serverless container hosting with readiness evaluations, migration needs, resource suggestions, and cost projections.
🌐 Networking
- Generally Available: Azure Front Door edge actions Azure Front Door now offers generally available edge actions, enabling serverless execution of lightweight JavaScript functions directly at the edge for real-time request and response customization.
- Generally Available: Azure Front Door mutual TLS authentication Azure Front Door now offers general availability for mutual TLS authentication, enabling organizations to secure B2B, IoT, and financial applications with X.509 client certificates.
- Public Preview: Azure Route Server dual-stack IPv4/IPv6 support Azure Route Server’s dual-stack IPv4/IPv6 support in public preview allows simultaneous route exchange for both address families over hybrid networks without disrupting existing IPv4 connections.
- Public Preview: Azure Front Door adds native gRPC support Azure Front Door introduces native gRPC support in public preview, allowing efficient routing to specific services and methods across Standard and Premium tiers.
- Public Preview: Rate limit grouping for Azure Front Door Premium WAF Azure adds rate limit grouping in Premium WAF for Front Door, enabling aggregated request counting per user to enhance DDoS defense.
- Public Preview: JWT validation with Azure Front Door Auth Policy Azure Front Door provides a public preview of JWT validation through its Auth Policy for real-time bearer token verification from Azure Active Directory at the edge.
- Building Private Connectivity for Fabric workloads The session showcases four private connectivity architectures for securely linking Microsoft Fabric and Power BI with on-premises SQL Server, tailored to various workload needs.
- Redefining Application Load Balancing in Azure Application Gateway Azure Application Gateway provides secure, scalable web app deployment with routing, SSL termination, WAF protection, up to 64 TCP connections per instance, and automatic scaling.
🔐 Security
- Securing AI Adoption in the Defense Industrial Base — Part 3: Governing AI and a 90-Day Plan This final part outlines a 90-day governance plan institutionalizing decision-making bodies, assigning owners and sponsors for agents, scaling oversight based on risk levels, and integrating ISO/IEC 42001 and RAISE framework principles to ensure repeatable, evidence-based control over AI systems in the Defense Industrial Base.
- FORGE Lab uses agentic tools to identify memory safety issues in Rust FORGE Lab used AI-powered Hyenas to uncover memory safety bugs in Rust's standard library, revealing three patterns where safe Rust can encounter undefined behavior due to unsafe code dependencies.
- Copilot Studio | Scoped for the Agent, Open to the URL Copilot Studio enforces agent-specific permissions client-side via the UseSpecificTools setting, handling authentication by the model instead of a custom MCP server, potentially causing unintended data access if misconfigured.
- Git security vulnerabilities announced Git has patched six vulnerabilities including arbitrary code execution during recursive clones, path traversal via malicious HEAD symrefs, stack corruption from malformed signatures, out-of-bounds reads in binary patches, heap overflows in large merges, and unsafe filesystem access through fast-import commands in the latest 2.56.1 release with fixes backported to older versions.
- Always On VPN Security Updates November 2026 The November 2026 updates patch two SSTP and one IKE vulnerabilities in Always On VPN while notably omitting any RRAS issues.
- Detecting sleeper agents hiding in AI model neurons Microsoft Research shows that fewer than 10 edited neurons can create a hidden backdoor in AI agents, detectable only under specific triggers, highlighting the necessity for independent monitoring of advanced models.
- Key Vault Certificate Expiry Emails: Certificate Contacts and the Issuance Policy You can set up Azure Key Vault to send expiry alerts via certificate contacts and the issuance policy, with an optional KQL alert for added security.
- Lessons from 59 Linux Kernel Bug Reports The FORGE Lab's three-month MDASH scan identified 92 issues in Linux kernel subsystems, including 8 root privilege-escalation exploits and 38 merged fixes.
- Microsoft Ignite 2026: New Sentinel Connectors Announcement The new Sentinel connectors at Microsoft Ignite 2026 integrate signals from over a dozen technologies to enhance investigation context and power the Integrated Security Operations Center with real-time threat visibility.
- How TÜV SÜD and Insight Enterprises are using Project Perception to investigate incidents Project Perception’s Attack Investigation Agent significantly reduces incident response time, lowering TÜV SÜD’s Kubernetes cluster issue investigation from two hours to 15 minutes and saving Insight Enterprises approximately 30 minutes through parallel analysis.
- The global socket: how threat actors are weaponizing GSocket for stealthy C2 GSocket allows threat actors to maintain persistent access on compromised Linux servers via a shared public relay network, avoiding fixed C2 addresses and traditional blocking.
- Extending confidential computing protection to your data disks Extending confidential computing protection to data disks enables guest-side encryption and attestation-based key release for attached volumes on Azure confidential VMs, enhancing workload security.
- Hands-On: Seculyze for Microsoft Sentinel – Part 2: Control Your Ingestion Cost Seculyze helps optimize ingestion costs in Azure Sentinel by allowing fine-grained control over data streams, potentially reducing expenses for workspaces with varied log volumes.
- Trained to betray: How post-training supply chain attacks hijack coding agents Microsoft Defender Research demonstrates how poisoned datasets and model weights can trigger unauthorized tool execution in coding agents with up to 94.8% success rates.
- SCKit abuses ScreenConnect to deploy remote access and target technician sessions The threat actor exploits ScreenConnect's extension vulnerabilities to hijack sessions and deploy SCKit for unauthorized remote access and malware distribution.
- Zero-Key Architecture on Azure: I Disabled Every Key Running Azure Speech, Storage, Functions and Application Insights with every access key switched off, authenticating with Entra ID only.
📦 Storage
- Public Preview: Modernize Arc-enabled Windows Servers with Azure File Sync Azure File Sync's public preview enables organizations to modernize Arc-enabled Windows Servers with Azure Files, providing hybrid access and a guided portal for seamless migration while preserving traditional file server operations.
- Public Preview: Distributed accelerator for cloud storage on AKS The Distributed Accelerator for Cloud Storage (DACS) boosts AI workload performance on AKS by caching data on local NVMe disks, reducing latency up to 70%.
- Azure NetApp Files: The Data Backbone for NVIDIA AI Model Training, Inference, and Microsoft Foundry Azure NetApp Files provides ultra-high throughput storage for NVIDIA's agentic AI stack and Microsoft Foundry, feeding GPU clusters during model training and supporting high-performance persistent storage for AI inference in Azure Kubernetes Service.
- Prevent Cross-Cloud Sync Data Loss with rclone on Linux rclone enables secure cross-cloud file synchronization on Linux, using checksums for data integrity and configurable bandwidth limits to prevent loss during migrations.
- Scale Azure Managed Lustre as your AI and HPC workloads grow Azure Managed Lustre now supports in-place capacity expansion for AI and HPC workloads, enabling growth without downtime or client interruptions while automatically scaling throughput.
- Azure Storage "This Request Is Not Authorized": Which Role Is Missing Maps each variant of the Azure Storage "not authorized" error to its cause and fix: Entra ID users with Owner or Contributor but no data role, and AuthorizationFailure, where the account firewall refused the address and no role assignment helps.
💻 Virtual Desktop Infrastructure
- Parallels RAS 22: what’s new in the latest release Parallels RAS 22 adds device posture checks, session recording, AVD Hybrid support, vSphere 9 integration, and improved file access management through the User Portal for stronger secure app delivery.
- Troubleshooting user connectivity issue with the help of Citrix Aidrien Citrix Aidrien helps administrators resolve VDA registration failures by identifying the specific Citrix Desktop Service error causing a launch failure.
🔗 Web
- Retirement: Support for .NET 8 and .NET 9 on Static Web Apps is ending November 10, 2026 Static Web Apps will cease supporting .NET 8 and .NET 9 on November 10, 2026, necessitating migration to .NET 10 for future security updates and support.
- Run on your schedule: App Service plan start/stop now in public preview The new App Service plan start/stop feature in Premium v4 and Premium memory-optimized v4 plans lets you pause or resume an entire environment’s workloads at once, simplifying management for development, test, and demo environments while ensuring all hosted apps are ready before traffic is restored.