What's New in Azure App Service at #MSBuild 2025
New App Service Premium v4 plan The new App Service Premium v4 (Pv4) plan has entered public preview at Microsoft Build 2025 for both Windows and Linux! This new plan is designed to support today's highly demanding application performance, scale, and budgets. Built on the latest "v6" general-purpose virtual machines and memory-optimized x64 Azure hardware with faster processors and NVMe temporary storage, it provides a noticeable performance uplift over prior generations of App Service Premium plans (over 25% in early testing). The Premium v4 offering includes nine new sizes ranging from P0v4 with a single virtual CPU and 4GB RAM all the way up through P5mv4, with 32 virtual CPUs and 256GB RAM, providing CPU and memory options to meet any business need. App Service Premium v4 plans provide attractive price-performance across the entire performance curve for both Windows and Linux customers. Premium v4 customers using pay-as-you-go (PAYG) on Azure App Service for Windows can expect to save up to 24% compared with prior Premium plans. We plan to provide deeper commitment-based discounts such as reserved instances and savings plan at GA. For more detailed pricing on the various CPU and memory options, see the pricing pages for Windows and Linux as well as the Azure Pricing Calculator. App Service currently has Pv4 deployed in a few regions with more regions being regularly added. For more details on how to configure app service plans with Premium v4 as well as a regularly updated list of regional availability, see the product documentation and start taking advantage of faster performance today! 2-zone Availability Zone support is now generally available With a recently completed platform update in May, customers now enjoy the 99.99% Availability Zone (AZ) SLA when running on only two instances (instead of three)! As part of this update more parts of the App Service footprint have enabled AZ support “in place”, which means many existing app service plans can now also use Availability Zones. Availability Zone configuration for app service plans is also now mutable. This means if an app service plan is running on an AZ-enabled part of the App Service footprint, customers can choose to enable and disable Availability Zone support at any time. Read more about the new Availability Zone options in the announcement article! ARM/CLI surface area for Availability Zone support has also been updated to provide increased visibility into AZ configuration details. The same enhanced visibility is also coming to the Azure Portal in June. With these changes customers can determine if an App Service plan is on an AZ-enabled scale unit, as well as how many zones are available for zone spanning. This allows customers to deploy with either two zones, or three zones (where available), of zone spanning for their App Service plans. For App Service plans that are AZ-enabled, customers will also be able to see the physical zone placement of each AZ enabled App Service plan. Availability Zone support is available on the new Premium v4 plan, and also supported with Premium v2, Premium v3, and the dedicated App Service Environment v3 (Isolated V2 plan). Check out the Availability Zone options for your App Service plans and start getting the benefits of zone resiliency today! .NET Aspire on Azure App Service .NET Aspire support is now available in public preview for App Service on Linux! .NET Aspire developers creating applications have an additional deployment option with App Service as a deployment target. Developers can create multi-app/multi-service .NET Aspire applications locally and deploy them into Azure using the new App Service deployment provider. The App Service and .NET Aspire teams worked together to create an App Service “provider” using .NET Aspire’s new “provider model”. The build provider translates the code-centric view of a .NET Aspire application topology into an Azure deployment mapped onto App Service constructs. The App Service provider supports securely deploying multiple .NET Aspire applications, with observability via the familiar .NET Aspire dashboard coming in the near future. The Getting Started with .NET Aspire on Azure App Service blog has instructions on how to create a .NET Aspire project for deployment onto App Service, as well a link for providing feedback. If you happen to be at Build 2025, drop by our booth or the theatre session “DEM548: How .NET Aspire on App Service enhances modern app development” to see live demonstrations of the App Service support for .NET Aspire! Using App Service to build agentic AI apps The last few months of intelligent app development have seen a frenetic pace of change with the rapid evolution of agents on Azure AI Foundry Agent Service and new agent extensibility options like Model Context Protocol (MCP) opening avenues for integrating existing data sources and APIs into agentic architectures. Here's a quick run-down of useful resources published recently: This article demonstrates hosting a remote MCP server on Azure App Service. The sample is an adaptation of the weather service example from the MCP site. The App Service variation also includes an azd template for easy experimentation via a CLI deployment to App Service! This article walks through integrating a .NET Core implementation of a “To-Do” list API running on App Service with an agent created on Azure AI Foundry Agent Service. It’s a straightforward example demonstrating how developers can bring together the power of AI agents with existing web API investments. Quick start guides for using App Service with Azure Open AI in your language of choice -- Python, Node, .NET, and Java. Using Microsoft Research’s latest 1-bit “super-small” language model, BitNet on App Service. Enhance search queries on text data stored in Azure SQL DB using natural language vector functions and Azure App Service. Includes an accompanying azd example. How to use Azure AI Search hybrid search capabilities from App Service with .NET (Blazor), Java (Spring Boot), Node (Express), or Python (FastAPI). Use GitHub Copilot to compare your application’s bicep against a representative “best practices” bicep definition and then generate the necessary bicep diff. In addition, using Sidecar for App Service on Linux, developers can easily connect Phi SLMs to their applications. Examples using the chat completion endpoint in the SLM sidecar extensions are available in this GitHub repo with code examples for .NET, Node, Python and Java. There are also accompanying docs for .NET, Node, Python (FastAPI) and Java (Spring Boot) which go into more details on using the SLM sidecar extensions. The sidecar extensions capability is also now enabled in the Azure Portal. AI Labs at Microsoft Build For those of you attending Microsoft Build in person, we will have labs for additional hands-on experience using AI with Azure App Service. LAB347: Add AI experiences to existing .NET apps using Sidecar in App Service This lab (first lab occurrence and second lab occurrence - see Exercise 4) covers an e-commerce inventory API (written in .NET) integrated with an agent running on Azure AI Foundry Agent Service. When a customer interacts with the AI agent it automatically invokes the appropriate web APIs to fetch real-time inventory information, add/remove products in a shopping cart, and increment/decrement product inventory. This is a great example of an AI powered agent grounded in a company’s ever changing transactional data. As a fun sidenote, GitHub Copilot was used extensively to build >95% of the sample application as well as to generate the OpenAPI specification that integrates the inventory web API with the AI agent! The same AI-on-App Service lab (Exercise 1) walks developers through integrating a basic Azure OpenAI chat interface into a web application. The lab also demonstrates using a background WebJob on Linux with Azure OpenAI (Exercise 2) to categorize user sentiment for product reviews. The lab also shows (Exercise 3) how to use a small language model (SLM) like Microsoft’s Phi-4 model in a WebJob to perform similar categorization, without the need to call out to an LLM. Although SLMs are not as powerful as LLMs, SLMs are an interesting alternative for integrating AI functionality where either cost, or control over AI data flows, are considerations. Azure SRE Agent for App Service One of the big announcements at Build this year was the Agentic DevOps announcement, which includes the new Azure SRE Agent. Designed to empower Site Reliability Engineers (SRE), the SRE Agent is a new agentic service that can manage Azure application platform services. including App Service, Functions, and Azure Container Apps to name just a few. It provides automatic incident response and mitigation, faster root cause analysis (RCA) of production issues, and continuous monitoring of application health and performance. With SRE Agent, you can use a natural language interface for managing your web applications on Azure App Service. To be an early adopter of the Agentic DevOps revolution, check out the announcement blog and sign-up to join the SRE Agent preview as it starts rolling out! WebJobs for App Service on Linux (GA) WebJobs for App Service on Linux just recently GA’d earlier this month. With this functionality developers can implement the same “infra-glue” style of background jobs that they have enjoyed with App service on Windows. Take a look at the documentationdemonstrating WebJobs support for shell scripts, Python, Java, .NET and Node on Linux! As mentioned earlier, the AI-on-App Service lab at this year’s Build conference has two code examples (see Exercise 2 and Exercise 3) demonstrating Linux WebJobs with Azure OpenAI, as well as a locally connected Phi-4 Small Language Model (SLM) sidecar, to categorize user sentiment for submitted product reviews. These are great examples of creatively using WebJobs to perform background batch-style work with your AI resources. Also keep an eye out for the upcoming WebJobs for Windows Containers GA which is planned “soon” this summer! Language and Framework Updates In addition to the release of .NET Aspire support for App Service, the App Service team has kept busy updating myriad Node, Python, Java/JBoss, .NET and PHP versions. To give an idea of the scope of effort keeping language and framework versions up to date across both Windows and Linux, App Service released more than two dozen language/framework specific updates in the last few weeks prior to Build. That represents the ongoing platform commitment to keeping languages regularly updated without the need for developers to explicitly invest time and effort doing so themselves. Just last month, Strapi support was introduced for App Service on Linux! Strapi is an open source headless Javascript based content management system that provides developers a robust platform for developing and delivering content across a variety of formats. The Azure Marketplace Strapi offering provides customization control, global availability and pre-built integration to essential Azure services like Azure Database for MySQL or PostgreSQL and Azure Email Communication Services. Deep dive on the details of hosting Strapi on App Service in this article. The custom error pages feature for App Service has also been updated just prior to Build. Custom error pages enable developers to customize the response rendered for common HTTP errors (403, 502 and 503) which are returned by the platform. This release includes a new option to always render custom errors regardless of whether the HTTP error was platform generated, or application generated. There will also be an Azure Portal update coming in June with support for the new custom error page features! Looking ahead to summer, stay tuned for the impending arrival of .NET 10 preview bits on App Service across both Windows and Linux! Networking and ASE Updates App Service support for public inbound IPv6 traffic is availablein most regions in public preview, with the service working towards a planned GA of inbound IPv6 support during the summer. Inbound IPv6 is supported for both IPv6-only upstream clients, as well dual-stack scenarios where a web application is reachable over either an IPv4 address or an IPv6 address. As part of an upcoming summer release, App Service will be delivering a public preview of *outbound* IPv6 traffic. For details on using IPv6 on App Service, as well to track all of the upcoming updates, consult this article: Announcing inbound IPv6 support in public preview - Azure App Service. For App Service Environment (ASE) customers, App Service will soon be releasing new support for adding custom Certificate Authorities (CAs) to an ASE. This new support will enable securing inbound TLS traffic using certificates issued by a custom Certificate Authority. Hybrid Connections customers will be happy to see that a new version of the App Service Hybrid Connection Manager (HCM) was just released just a few weeks ago. The new HCM delivers updated UX support for both Linux and Windows customers, enhanced logging and connection testing, and a brand new CLI for scripting and command-line management of Hybrid Connections! You might have missed it, but there was a recent addition to the troubleshooting options on App Service with the new Network Troubleshooter! The Network Troubleshooter offers comprehensive analysis and actionable insights to resolve connectivity failures for both Linux and Windows web apps. It tests connectivity to Azure resources like Storage, Redis, SQL Server, MySQL server, and other apps running on App Service. It diagnoses connectivity problems with Private endpoints, Service endpoints, and Internet-based endpoints, detects NAT gateways, and investigates DNS failures with custom DNS servers. Additionally, it provides actionable recommendations and surfaces any network rules it finds that are blocking connectivity. If you regularly wrestle with connectivity challenges, give the Network Troubleshooter a try! Next Steps Developers can learn more about Azure App Service at Getting Started with Azure App Service. Stay up to date on new features and innovations on Azure App Service via Azure Updates as well as the Azure App Service (@AzAppService) X feed. There is always a steady stream of great deep-dive technical articles about App Service as well as the breadth of developer focused Azure services over on the Apps on Azure blog. And lastly take a look at Azure App Service Community Standups hosted on the Microsoft Azure Developers YouTube channel. The Azure App Service Community Standup series regularly features walkthroughs of new and upcoming features from folks that work directly on the product! Build 2025 Session Reference (Note: all times below are listed in Seattle time - Pacific Daylight Time) (Note: some labs have more than one timeslot spanning multiple days) Innovate, deploy, & optimize your apps without infrastructure hassles https://build.microsoft.com/en-US/sessions/BRK201 Monday, May 19 th 11:15 AM – 12:15 PM Pacific Daylight Time Arch, 705 Pike, Level 6, Room 606 Breakout, Streaming Online and Recorded Session (BRK201) Quickly build, deploy, and scale web apps and APIs globally with App Service https://build.microsoft.com/en-US/sessions/BRK200 Tuesday, May 20 th 11:45 AM – 12:45 PM Pacific Daylight Time Arch, 705 Pike, Level 6, Room 608 Breakout, Streaming Online and Recorded Session (BRK200) Simplifying .NET upgrades with GitHub Copilot https://build.microsoft.com/en-US/sessions/DEM549 Monday, May 19 th 5:05 PM - 5:20 PM Pacific Daylight Time Arch, 705 Pike, Level 4, Hub, Theater B Demo Session – Also Recorded (DEM549) Use Azure SRE Agent to automate tasks and increase site reliability https://build.microsoft.com/en-US/sessions/DEM550 Tuesday, May 20 th 5:10 PM - 5:25 PM Pacific Daylight Time Arch, 705 Pike, Level 4, Hub, Theater A Demo Session – Also Recorded (DEM550) How .NET Aspire on App Service enhances modern app development https://build.microsoft.com/en-US/sessions/DEM548 Wednesday, May 21 st 2:00 PM - 2:15 PM Pacific Daylight Time Arch, 705 Pike, Level 4, Hub, Theater B Demo Session – Also Recorded (DEM548) Add AI experiences to existing .NET apps using Sidecars in App Service [Note: Lab participants will be able to try Phi-4 and Azure AI Foundry Agent service scenarios in this lab.] https://build.microsoft.com/en-US/sessions/LAB347 Monday, May 19 th 4:45 PM - 6:00 PM Pacific Daylight Time Arch, 800 Pike, Level 1, Yakima 1 Hands on Lab – In-Person Only (LAB347) You can also work through the lab with your own Azure subscription! Code is available at https://github.com/Azure-Samples/Build2025-LAB347. Deploy the lab resources using the included resource provisioning template (https://github.com/Azure-Samples/Build2025-LAB347/blob/main/resources/lab347.json). You can deploy the template by searching on “Deploy a custom template” in the Azure Portal, and copying and pasting the template into the “Build your own template in the editor option”! Add AI experiences to existing .NET apps using Sidecars in App Service [Note: Lab participants will be able to try Phi-4 and Azure AI Foundry Agent service scenarios in this lab.] https://build.microsoft.com/en-US/sessions/LAB347-R1 Wednesday, May 21 st 4:30 PM - 5:45 PM Pacific Daylight Time Arch, 800 Pike, Lower Level, Skagit 5 Hands on Lab – In-Person Only (LAB347-R1) You can also work through the lab with your own Azure subscription! Code is available at https://github.com/Azure-Samples/Build2025-LAB347. Deploy the lab resources using the included resource provisioning template (https://github.com/Azure-Samples/Build2025-LAB347/blob/main/resources/lab347.json). You can deploy the template by searching on “Deploy a custom template” in the Azure Portal, and copying and pasting the template into the “Build your own template in the editor option”! Modernizing .NET Applications using Azure Migrate and GitHub Copilot https://build.microsoft.com/en-US/sessions/LAB343 Tuesday, May 20 th 5:15 PM - 6:30 PM Pacific Daylight Time Arch, 800 Pike, Level 1, Yakima 1 Hands on Lab – In-Person Only (LAB343) Modernizing .NET Applications using Azure Migrate and GitHub Copilot https://build.microsoft.com/en-US/sessions/LAB343-R1 Thursday, May 22 nd 10:15 AM – 11:30 AM Pacific Daylight Time Arch, 800 Pike, Level 2, Chelan 2 Hands on Lab – In-Person Only (LAB343-R1)Building the Agentic Future
As a business built by developers, for developers, Microsoft has spent decades making it faster, easier and more exciting to create great software. And developers everywhere have turned everything from BASIC and the .NET Framework, to Azure, VS Code, GitHub and more into the digital world we all live in today. But nothing compares to what’s on the horizon as agentic AI redefines both how we build and the apps we’re building. In fact, the promise of agentic AI is so strong that market forecasts predict we’re on track to reach 1.3 billion AI Agents by 2028. Our own data, from 1,500 organizations around the world, shows agent capabilities have jumped as a driver for AI applications from near last to a top three priority when comparing deployments earlier this year to applications being defined today. Of those organizations building AI agents, 41% chose Microsoft to build and run their solutions, significantly more than any other vendor. But within software development the opportunity is even greater, with approximately 50% of businesses intending to incorporate agentic AI into software engineering this year alone. Developers face a fascinating yet challenging world of complex agent workflows, a constant pipeline of new models, new security and governance requirements, and the continued pressure to deliver value from AI, fast, all while contending with decades of legacy applications and technical debt. This week at Microsoft Build, you can see how we’re making this future a reality with new AI-native developer practices and experiences, by extending the value of AI across the entire software lifecycle, and by bringing critical AI, data, and toolchain services directly to the hands of developers, in the most popular developer tools in the world. Agentic DevOps AI has already transformed the way we code, with 15 million developers using GitHub Copilot today to build faster. But coding is only a fraction of the developer’s time. Extending agents across the entire software lifecycle, means developers can move faster from idea to production, boost code quality, and strengthen security, while removing the burden of low value, routine, time consuming tasks. We can even address decades of technical debt and keep apps running smoothly in production. This is the foundation of agentic DevOps—the next evolution of DevOps, reimagined for a world where intelligent agents collaborate with developer teams and with each other. Agents introduced today across GitHub Copilot and Azure operate like a member of your development team, automating and optimizing every stage of the software lifecycle, from performing code reviews, and writing tests to fixing defects and building entire specs. Copilot can even collaborate with other agents to complete complex tasks like resolving production issues. Developers stay at the center of innovation, orchestrating agents for the mundane while focusing their energy on the work that matters most. Customers like EY are already seeing the impact: “The coding agent in GitHub Copilot is opening up doors for each developer to have their own team, all working in parallel to amplify their work. Now we're able to assign tasks that would typically detract from deeper, more complex work, freeing up several hours for focus time." - James Zabinski, DevEx Lead at EY You can learn more about agentic DevOps and the new capabilities announced today from Amanda Silver, Corporate Vice President of Product, Microsoft Developer Division, and Mario Rodriguez, Chief Product Office at GitHub. And be sure to read more from GitHub CEO Thomas Dohmke about the latest with GitHub Copilot. At Microsoft Build, see agentic DevOps in action in the following sessions, available both in-person May 19 - 22 in Seattle and on-demand: BRK100: Reimagining Software Development and DevOps with Agentic AI BRK 113: The Agent Awakens: Collaborative Development with GitHub Copilot BRK118: Accelerate Azure Development with GitHub Copilot, VS Code & AI BRK131: Java App Modernization Simplified with AI BRK102: Agent Mode in Action: AI Coding with Vibe and Spec-Driven Flows BRK101: The Future of .NET App Modernization Streamlined with AI New AI Toolchain Integrations Beyond these new agentic capabilities, we’re also releasing new integrations that bring key services directly to the tools developers are already using. From the 150 million GitHub users to the 50 million monthly users of the VS Code family, we’re making it easier for developers everywhere to build AI apps. If GitHub Copilot changed how we write code, Azure AI Foundry is changing what we can build. And the combination of the two is incredibly powerful. Now we’re bringing leading models from Azure AI Foundry directly into your GitHub experience and workflow, with a new native integration. GitHub models lets you experiment with leading models from OpenAI, Meta, Cohere, Microsoft, Mistral and more. Test and compare performance while building models directly into your codebase all within in GitHub. You can easily select the best model performance and price side by side and swap models with a simple, unified API. And keeping with our enterprise commitment, teams can set guardrails so model selection is secure, responsible, and in line with your team’s policies. Meanwhile, new Azure Native Integrations gives developers seamless access to a curated set of 20 software services from DataDog, New Relic, Pinecone, Pure Storage Cloud and more, directly through Azure portal, SDK, and CLI. With Azure Native Integrations, developers get the flexibility to work with their preferred vendors across the AI toolchain with simplified single sign-on and management, while staying in Azure. Today, we are pleased to announce the addition of even more developer services: Arize AI: Arize’s platform provides essential tooling for AI and agent evaluation, experimentation, and observability at scale. With Arize, developers can easily optimize AI applications through tools for tracing, prompt engineering, dataset curation, and automated evaluations. Learn more. LambdaTest HyperExecute: LambdaTest HyperExecute is an AI-native test execution platform designed to accelerate software testing. It enables developers and testers to run tests up to 70% faster than traditional cloud grids by optimizing test orchestration, observability and streamlining TestOps to expedite release cycles. Learn more. Mistral: Mistral and Microsoft announced a partnership today, which includes integrating Mistral La Plateforme as part of Azure Native Integrations. Mistral La Plateforme provides pay-as-you-go API access to Mistral AI's latest large language models for text generation, embeddings, and function calling. Developers can use this AI platform to build AI-powered applications with retrieval-augmented generation (RAG), fine-tune models for domain-specific tasks, and integrate AI agents into enterprise workflows. MongoDB (Public Preview): MongoDB Atlas is a fully managed cloud database that provides scalability, security, and multi-cloud support for modern applications. Developers can use it to store and search vector embeddings, implement retrieval-augmented generation (RAG), and build AI-powered search and recommendation systems. Learn more. Neon: Neon Serverless Postgres is a fully managed, autoscaling PostgreSQL database designed for instant provisioning, cost efficiency, and AI-native workloads. Developers can use it to rapidly spin up databases for AI agents, store vector embeddings with pgvector, and scale AI applications seamlessly. Learn more. Java and .Net App Modernization Shipping to production isn’t the finish line—and maintaining legacy code shouldn’t slow you down. Today we’re announcing comprehensive resources to help you successfully plan and execute app modernization initiatives, along with new agents in GitHub Copilot to help you modernize at scale, in a fraction of the time. In fact, customers like Ford China are seeing breakthrough results, reducing up to 70% of their Java migration efforts by using GitHub Copilot to automate middleware code migration tasks. Microsoft’s App Modernization Guidance applies decades of enterprise apps experience to help you analyze production apps and prioritize modernization efforts, while applying best practices and technical patterns to ensure success. And now GitHub Copilot transforms the modernization process, handling code assessments, dependency updates, and remediation across your production Java and .NET apps (support for mainframe environments is coming soon!). It generates and executes update plans automatically, while giving you full visibility, control, and a clear summary of changes. You can even raise modernization tasks in GitHub Issues from our proven service Azure Migrate to assign to developer teams. Your apps are more secure, maintainable, and cost-efficient, faster than ever. Learn how we’re reimagining app modernization for the era of AI with the new App Modernization Guidance and the modernization agent in GitHub Copilot to help you modernize your complete app estate. Scaling AI Apps and Agents Sophisticated apps and agents need an equally powerful runtime. And today we’re advancing our complete portfolio, from serverless with Azure Functions and Azure Container Apps, to the control and scale of Azure Kubernetes Service. At Build we’re simplifying how you deploy, test, and operate open-source and custom models on Kubernetes through Kubernetes AI Toolchain Operator (KAITO), making it easy to inference AI models with the flexibility, auto-scaling, pay-per-second pricing, and governance of Azure Container Apps serverless GPU, helping you create real-time, event-driven workflows for AI agents by integrating Azure Functions with Azure AI Foundry Agent Service, and much, much more. The platform you choose to scale your apps has never been more important. With new integrations with Azure AI Foundry, advanced automation that reduces developer overhead, and simplified operations, security and governance, Azure’s app platform can help you deliver the sophisticated, secure AI apps your business demands. To see the full slate of innovations across the app platform, check out: Powering the Next Generation of AI Apps and Agents on the Azure Application Platform Tools that keep pace with how you need to build This week we’re also introducing new enhancements to our tooling to help you build as fast as possible and explore what’s next with AI, all directly from your editor. GitHub Copilot for Azure brings Azure-specific tools into agent mode in VS Code, keeping you in the flow as you create, manage, and troubleshoot cloud apps. Meanwhile the Azure Tools for VS Code extension pack brings everything you need to build apps on Azure using GitHub Copilot to VS Code, making it easy to discover and interact with cloud services that power your applications. Microsoft’s gallery of AI App Templates continues to expand, helping you rapidly move from concept to production app, deployed on Azure. Each template includes fully working applications, complete with app code, AI features, infrastructure as code (IaC), configurable CI/CD pipelines with GitHub Actions, along with an application architecture, ready to deploy to Azure. These templates reflect the most common patterns and use cases we see across our AI customers, from getting started with AI agents to building GenAI chat experiences with your enterprise data and helping you learn how to use best practices such as keyless authentication. Learn more by reading the latest on Build Apps and Agents with Visual Studio Code and Azure Building the agentic future The emergence of agentic DevOps, the new wave of development powered by GitHub Copilot and new services launching across Microsoft Build will be transformative. But just as we’ve seen over the first 50 years of Microsoft’s history, the real impact will come from the global community of developers. You all have the power to turn these tools and platforms into advanced AI apps and agents that make every business move faster, operate more intelligently and innovate in ways that were previously impossible. Learn more and get started with GitHub CopilotReimagining App Modernization for the Era of AI
This blog highlights the key announcements and innovations from Microsoft Build 2025. It focuses on how AI is transforming the software development lifecycle, particularly in app modernization. Key topics include the use of GitHub Copilot for accelerating development and modernization, the introduction of Azure SRE agent for managing production systems, and the launch of the App Modernization Guidance to help organizations modernize their applications with AI-first design. The blog emphasizes the strategic approach to modernization, aiming to reduce complexity, improve agility, and deliver measurable business outcomes
FSI Knowledge Mining and Intelligent Document Process Reference Architecture
FSI customers such as insurance companies and banks rely on their vast amounts of data to provide sometimes hundreds of individual products to their customers. From assessing product suitability, underwriting, fraud investigations, and claims handling, many employees and applications depend on accessing this data to do their jobs efficiently. Since the capabilities of GenAI have been realised, we have been helping our customers in this market transform their business with unified systems that simplify access to this data and speed up the processing times of these core tasks, while remaining compliant with the numerous regulations that govern the FSI space. Combining the use of Knowledge Mining with Intelligent Document processing provides a powerful solution to reduce the manual effort and inefficacies of ensuring data integrity and retrieval across the many use cases that most of our customers face daily. What is Knowledge Mining and Intelligent Document Processing? Knowledge Mining is a process that transforms large, unstructured data sets into searchable knowledge stores. Traditional search methods often rely on keyword matching, which can miss the context of the information. In contrast, knowledge mining uses advanced techniques like natural language processing (NLP) to understand the context and meaning behind the data, providing a robust searching mechanism that can look across all these data sources, understand the relationships between the data therefore providing more accurate and relevant results. Intelligent Document Processing (IDP) is a workflow automation technology designed to scan, read, extract, categorise, and organise meaningful information from large streams of data. Its primary function is to extract valuable information from extensive data sets without human input, thereby increasing processing speed and accuracy while reducing costs. By leveraging a combination of Artificial Intelligence (AI), Machine Learning (ML), Optical Character Recognition (OCR), and Natural Language Processing (NLP), IDP handles both structured and unstructured documents. By ensuring that the processed data meets the "gold standard" - structured, complete, and compliant - IDP helps organizations maintain high-quality, reliable, and actionable data. The Power of Knowledge Mining and Intelligent Document Processing as a Unified Solution Knowledge Mining excels at quickly responding to natural language queries, providing valuable insights and making previously unsearchable data accessible. At the same time, IDP ensures that the processed data meets the "gold standard"—structured, complete, and compliant—making it both reliable and actionable. Together, these technologies empower organisations to harness the full potential of their data, driving better decision-making and improved efficiency. __________________________________________________________________ Meet Alex: A Day in the Life of a Fraud Case Worker Responsibilities: Investigate potential fraud cases by manually searching across multiple systems. Read and analyse large volumes of information to filter out relevant data. Ensure compliance with regulatory requirements and maintain data accuracy. Prepare detailed reports on findings and recommendations. Lost in Data: The Struggles of Manual Fraud Investigation Alex receives a new fraud case and starts by manually searching through multiple systems to gather information. This process takes several hours, and Alex has to read through numerous documents and emails to filter out relevant data. The inconsistent data formats and locations make it challenging to ensure accuracy. By the end of the day, Alex is exhausted and has only made limited progress on the case. Effortless Efficiency: Fraud Investigation Transformed with Knowledge Mining and IDP Alex receives a new fraud case and needs to gather all relevant information quickly. Instead of manually searching through multiple systems, Alex inputs the following natural language query into the unified system: "Show me all documents, emails, and notes related to the recent transactions of client X that might indicate fraudulent activity." The system quickly retrieves and presents a comprehensive summary of all relevant documents, emails, and notes, ensuring that the data is structured, complete, and compliant. This allows Alex to focus on analysing the data and making informed decisions, significantly improving the efficiency and accuracy of the investigation. How has Knowledge Mining and IDP transformed Alex's role? Before implementing Knowledge Mining and Intelligent Document Processing, Alex faced a manual process of searching across multiple systems to gather information. This was time-consuming and labour-intensive, often leading to delays in investigations. The overwhelming volume of data from various sources made it difficult to filter out relevant information, and the inconsistent data formats and locations increased the risk of errors. This high workload not only reduced Alex's efficiency but also led to burnout and decreased job satisfaction. However, with the introduction of a unified system powered by Knowledge Mining and IDP, these challenges were significantly mitigated. Automated searches using natural language queries allowed Alex to quickly find relevant information, while IDP ensured that the data processed was structured, complete, and compliant. This unified system provided a comprehensive view of the data, enabling Alex to make more informed decisions and focus on higher-value tasks, ultimately improving productivity and job satisfaction. ____________________________________________________________________ Example Architecture Knowledge Mining Users can interact with the system through a portal on the customer’s front-end of choice. This will serve as the entry point for submitting queries and accessing the knowledge mining service. Front-end options could include web apps, container services or serverless integrations. Azure AI Search provides powerful RAG capabilities. Meanwhile, Azure Open AI provides access to large language models to summarise responses. These services combined will take the user’s query to search the knowledge base and return relevant information which can be augmented as required. Prompt engineering can provide customisation to how the data is returned. You define what the data sources your Azure AI Search will consume. This can be Azure storage services or other data repositories. Data that meets a pre-defined gold standard is queried by Azure AI Search and relevant data is returned to the user. Gold standard data could be based on compliance or business needs. Power BI can be used to create analytical reports based on the data retrieved and processed. This step involves visualising the data in an interactive and user-friendly manner, allowing users to gain insights and make data-driven decisions. Intelligent Document Processing (Optional) Azure Data Factory is a data integration service that allows you to create workflows for data movement and transforming data at scale. This business data can be easily ingested to your Azure data storage solutions using pre-built connectors. This event driven approach ensures that as new data is generated, it can automatically be processed and ready for use in your knowledge mining solution. Data can be transformed using Functions apps and Azure OpenAI. Through prompt engineering, the large language model (LLM) can highlight specific issues in the documents, such as grammatical errors, irrelevant content, or incomplete information. The LLM can then be used to rewrite text to improve clarity and accuracy, add missing information, or reformat content to adhere to guidelines. Transformed data is stored as gold standard data. ____________________________________________________________________ Additional Cloud Considerations Networking VNETs (Virtual Networks) are a fundamental component of cloud infrastructure that enable secure and isolated networking configurations within a cloud environment. They allow different resources, such as virtual machines, databases, and services, to communicate with each other securely. Virtual networks ensure that services such as Azure AI Search, Azure OpenAI, and Power BI, can securely communicate with each other. This is crucial for maintaining the integrity and confidentiality of sensitive financial data. Express Route or VPN are expected to be used when connecting on-premises infrastructure to Azure for several reasons. Your company Azure ExpressRoute provides a private, reliable, and high-speed connection between your data center and Microsoft Azure. It allows you to extend your infrastructure to Azure by providing private access to resources deployed in Azure Virtual Networks and public services like App service, private end points to various other services. This private peering ensures that your traffic never enters the public Internet, enhancing security and performance. ExpressRoute uses Border Gateway Protocol (BGP) for dynamic routing between your on-premises networks and Azure, ensuring efficient and secure data exchange. It also offers built-in redundancy and high availability, making it a robust solution for critical workloads. Azure Front Door is a cloud-based Content Delivery Network (CDN) and application delivery service provided by Microsoft. It offers several key features, including global load balancing, dynamic site acceleration, SSL offloading, and a web application firewall, making it an ideal solution for optimizing and protecting web applications. We are expecting to use Front door in scenarios when the architecture will be expected to be used by users outside the organisation. Azure API Management in this scenario is expected to be used when we look to rollout the solution to larger groups. We look to then integrate much more security, rate limiting, load balancing, etc. Monitoring and Governance Azure Monitor: This service collects and analyses telemetry data from various resources, providing insights into the performance and health of the system. It enables proactive identification and resolution of issues, ensuring the system runs smoothly. Azure Cost Management and Billing: Provides tools for monitoring and controlling costs associated with the solution. It offers insights into spending patterns and resource usage, enabling efficient financial governance. Application Insights: Provides application performance monitoring (APM) designed to help you understand how your applications are performing and to identify issues that may affect their performance and reliability These components together ensure that the Knowledge Mining and Intelligent Document Processing solution is monitored for performance, secured against threats, compliant with regulations, and managed efficiently from a cost perspective. ____________________________________________________________________ Next steps: Identify the data and its sources that will feed into your own Knowledge Mine. Consider if you also need to implement Intelligent Document Processing to ensure data quality. Define your 'gold standards'. These guidelines will determine how your data might be transformed. Consider how to provide access to the data through an application portal, choose the right front-end technology for your use case. Once you have configured Azure AI search to point to the chosen data, consider how you might augment responses using Azure AI LLM models. Useful resources AI Landing Zone reference architecture Azure and Open AI with API Manager Secure connectivity from on premesis to Azure hosted solutionsAzure Kubernetes Fleet Manager Demo with Terraform Code
Introduction Azure Kubernetes Fleet Manager (Fleet Manager) simplifies the at-scale management of multiple Azure Kubernetes Service (AKS) clusters by treating them as a coordinated “fleet.” One Fleet Manager hub can manage up to 100 AKS clusters in a single Azure AD tenant and region scope, so you can register, organize, and operate a large number of clusters from a single control plane. In this walkthrough, we’ll explore: The key benefits and considerations of using Fleet Manager A real-world e-commerce use case How to deploy a Fleet Manager hub, AKS clusters, and Azure Front Door with Terraform How everything looks and works in the Azure portal Along the way, you’ll see screenshots from my demo environment to illustrate each feature. Why Use Fleet Manager? Managing dozens or even hundreds of AKS clusters individually quickly become unmanageable. Fleet Manager introduces: Centralized control plane: Register AKS clusters across subscriptions/regions under one fleet. Orchestrated upgrades: Define update runs, stages, and groups (ring-based rollouts). Resource propagation: Declaratively push Kubernetes objects (Namespaces, RBAC, ConfigMaps) from hub → members. Cross-cluster L4 load balancing (preview): Distribute TCP/UDP traffic across clusters for high availability. Auto-upgrade profiles: Automatically keep clusters up to date with minimal manual effort. Portal Walkthrough: Exploring Your Fleet Once your Fleet Manager hub and member clusters are up, here’s how it looks in the Azure portal. Member Clusters The Member clusters blade shows all onboarded clusters, their membership status, update group assignment, and Kubernetes version. Figure: Four clusters (two dev, two prod) successfully joined to the fleet, all running version 1.32.3. Multi-Cluster Update Under multi-cluster update, you can manage both Auto-upgrade profiles and Strategies. Auto-upgrade profiles let you enable continuous updates by channel (e.g., Stable) and node image: Strategies define how clusters are grouped and staged during an update run: Figure: We’ve created development-auto-upgrade and production-auto-upgrade profiles, along with matching strategies. Fleet Overview Back on the hub’s Overview blade, you get at-a-glance insights: total member clusters, Kubernetes versions spread, and node image versions. Figure: The hub reports 4 member clusters (all on 1.32.3), and the node pools all share the same image version. Azure Front Door Origin Groups To demonstrate multi-cluster traffic routing, our Terraform deploy includes an Azure Front Door profile with two origin groups (dev & prod). Here’s the Origin groups blade: And the Front Door Overview, showing the endpoint hostname and associated origin groups: Figure: Front Door is configured to route /dev/* to the dev clusters and /prod/* to the prod clusters via these origin groups. Benefits & Considerations Benefits One pane of glass for up to 100 AKS clusters. Ring-based upgrades minimize risk with staged rollouts. Declarative propagation of configs and policies. Global traffic distribution at TCP/UDP (L4) level. Extensible roadmap: Arc support, region failover, Terraform enhancements. Considerations Hub is management-only: No user workloads on the hub. 100-cluster limit per fleet. Regional scope: Hub deployed in one region, though it can manage clusters anywhere. Private hub networking: Private AKS hub requires VNet/jumpbox connectivity. Preview features: Multi-cluster L4 load balancing and Terraform support for update groups are still in preview. Real-World Use Case: Global E-Commerce A multinational retailer runs dev & prod AKS clusters in North America and Europe. They needed: Consistent feature flags & RBAC across clusters Safe, staged upgrades (dev → prod) High-availability checkout traffic routed to healthy clusters Solution with Fleet Manager: Onboard all four clusters into one fleet. Propagate feature-toggle ConfigMaps and RBAC from hub to members. Define update strategies for dev and prod, then run upgrades via CLI or portal. Use Azure Front Door for global routing, failing over between regions. They cut upgrade windows by 60%, eliminated manual sync tasks, and boosted resilience. Reference Architecture for Demo: Deployment with Terraform All of the above is automated in the aks-fleet-manager GitHub repo. Here’s a quick start: 1. Clone repo git clone https://github.com/saswatmohanty01/aks-fleet-manager.git cd aks-fleet-manager/terraform 2. Install CLI tools chmod +x ../scripts/setup-cli.sh ../scripts/setup-cli.sh 3. Authenticate & select subscription az login az account set -s <subscription-id> 4. Initialize Terraform terraform init 5. Configure variables (terraform.tfvars): primary_region = "eastus" secondary_region = "westeurope" resource_prefix = "mycompany" dev_node_count = 2 prod_node_count = 3 6. Plan & apply terraform plan -out=tfplan terraform apply tfplan 7. Create update groups (post-deploy) cd ../scripts chmod +x create-update-groups.sh ./create-update-groups.sh Once complete (about 10–15 minutes), you’ll have: 4 AKS clusters (dev/prod in two regions) A Fleet Manager hub with 4 member clusters Auto-upgrade profiles and strategies An Azure Front Door endpoint routing /dev/ and /prod/ Known Issue. Manual Step in Azure Front Door Refer GitHub: README.md Get the terraform output for all four AKS clusters service endpoint IP addresses. You can get it from step 3 using kubectl get svc for all four clusters. There is a bug, which does not allow to update the service IP addresses for each AKS cluster in Azure Frontdoor->Origin Groups Manually update the IP addresses for Dev and Prod AKS cluster service IP addresses. Go to Azure portal->Azure Front door->Settings->Origin Groups->dev-origin-group Manually update the IP addresses for Dev and Prod AKS cluster service IP addresses. Go to Azure portal->Azure Front door->Settings->Origin Groups->prod-origin-group VS Code Experience Follow the VsCode Deployment Guide from GitHub Repo Conclusion & Next Steps Azure Kubernetes Fleet Manager reduces the pain of managing multi-cluster AKS environments by centralizing control, orchestrating upgrades, and enabling global traffic patterns. To go further: Experiment with auto-upgrade profiles to automate patch deployments. Integrate Fleet operations into CI/CD pipelines with az fleet CLI or Terraform (as features mature). Explore GitOps workflows (Flux/Argo CD) for multi-cluster app deployments. Fleet Manager is evolving rapidly—keep an eye on the preview features and Terraform provider updates. With Fleet Manager, managing up to 100 AKS clusters doesn’t have to be a headache. Give it a try and share your experiences! References Azure Kubernetes Fleet Manager overview (Microsoft Learn) QuickStart: Create a fleet and join member clusters (Microsoft Learn) Fleet Manager CLI commands (Azure CLI docs) aks-fleet-manager GitHub repo & docs Architecture diagram: architecture-diagrams Happy clustering!Get Ready for .NET Conf: Focus on Modernization
We’re excited to announce the topics and speakers for .NET Conf: Focus on Modernization, our latest virtual event on April 22-23, 2025! This event features live sessions from .NET and cloud computing experts, providing attendees with the latest insights into modernizing .NET applications, including technical upgrades, cloud migration, and tooling advancements. To get ready, visit the .NET Conf: Focus on Modernization home page and click Add to Calendar so you can save the date on your calendar. From this page, on the day of the event you’ll be able to join a live stream on YouTube and Twitch. We will also make the source code for the demos available on GitHub and the on-demand replays will be available on our YouTube channel. Learn more: https://focus.dotnetconf.net/ Why attend? In the fast-changing technological environment we now find ourselves, it has never been more urgent to modernize enterprise .NET applications to maintain competitiveness and stay ahead of the next innovation. Updating .NET applications for the cloud is a major business priority and involves not only technical upgrades and cloud migration, but also improvements in tooling, processes, and skills. At this event, you will get the end to end insights across latest tools, innovations, and best practices for successful .NET modernization. What can developers expect? The event will run live for up to five hours each day, covering different aspects of .NET modernizations. Scott Hanselman will set the tone for day one with discussion of the experiences and processes to modernize .NET applications in the era of AI. This will be followed by expert sessions on upgrading .NET apps and modernizing both your apps and data to the cloud. Day two will soar higher into the clouds, with sessions to help with cloud migration, cloud development, and infusing AI into your apps. You can interact with experts and ask questions to deepen your expertise, as we broadcast live on YouTube, or Twitch. Recordings of all sessions will be available with materials after the event. Agenda Here’s a quick snapshot of the schedule. Things may change, and we recommend that you please visit the event home page for the latest agenda and session times: https://focus.dotnetconf.net/agenda Day 1 – April 22, Tuesday Time (PDT) Session 8:00 am Modernizing .NET: Future-ready applications in the era of AI Scott Hanselman, Chet Husk, McKenna Barlow 9:00 am Deep dive into the upcoming AI-assisted tooling to upgrade .NET apps Chet Husk, McKenna Barlow 10:00 am Use Reliable Web App patterns to confidently replatform your web apps Pablo Lopes 11:00 am Modernize Data-Driven Apps (No AI Needed) Jerry Nixon 12:00 pm Modernize from ASP.NET to ASP.NET Core: The Future is Now Taylor Southwick Day 2 – April 23, Wednesday Time (PDT) Session 8:00 am Unblock .NET modernization with AI-assisted app and code assessment tools Michael Yen-Chi Ho 9:00 am Cloud development doesn't have to be painful thanks to .NET Aspire Maddy Montaquila (Leger) 10:00 am Introducing Artificial Intelligence to your application Jordan Matthiesen 11:00 am Modernizing your desktop: From WinForms to Blazor, Azure, and AI Santiago Arango Toro Save the Date! .NET Conf: Focus on Modernization is a free, two-day livestream event that you won’t want to miss. Tune in on April 22 and 23, 2025, ask questions live, and learn how to get your .NET applications ready for the AI revolution. Save the date! Stay tuned for more updates and detailed session information. We can’t wait to see you there!Code the Future with Java and AI – Join Me at JDConf 2025
JDConf 2025 is just around the corner, and whether you’re a Java developer, architect, team leader, or decision maker I hope you’ll join me as we explore how Java is evolving with the power of AI and how you can start building the next generation of intelligent applications today. Why JDConf 2025? With over 22 expert-led sessions and 10+ hours of live content, JDConf is packed with learning, hands-on demos, and real-world solutions. You’ll hear from Java leaders and engineers on everything from modern application design to bringing AI into your Java stack. It’s free, virtual and your chance to connect from wherever you are. (On-demand sessions will also be available globally from April 9–10, so you can tune in anytime from anywhere.) Bring AI into Java Apps At JDConf 2025, we are going beyond buzzwords. We’ll show you how to bring AI into real Java apps, using patterns and tools that work today. First, we’ll cover Retrieval-Augmented Generation (RAG), a design pattern where your app retrieves the right business data in real time, and combines it with AI models to generate smart, context-aware responses. Whether it is answering support queries, optimizing schedules, or generating insights, RAG enables your app to think in real time. Second, we’ll introduce AI agents -- software entities that do more than respond. They act. Think about automating production line scheduling at an auto manufacturer or rebooking delayed flights for passengers. These agents interact with APIs, reason over data, and make decisions, all without human intervention. Third, we’ll explore the complete AI application platform on Azure. It is built to work with the tools Java developers already know - from Spring Boot to Quarkus - and includes OpenAI and many other models, vector search with PostgreSQL, and libraries like Spring AI and LangChain4j. Here are just two example stacks: Spring Boot AI Stack: any app hosting services like Azure Container Apps or App Service + Spring AI + OpenAI + PostgreSQL for business data and vector data store. Quarkus AI Stack: any app hosting services like Azure Container Apps or App Service + LangChain4j + OpenAI + PostgreSQL for business data and vector data store. This is how you turn existing Java apps into intelligent, interactive systems, without reinventing everything. Whether you are an experienced developer or just starting out, JDConf offers valuable opportunities to explore the latest advancements in Java, cloud, and AI technologies; gain practical insights; and connect with Java experts from across the globe – including Java 25, Virtual Threads, Spring Boot, Jakarta EE 12, AI developer experiences, Spring AI, LangChain4j, combining data and AI, automated refactoring to Java app code modernization. We’ll also show you how GitHub Copilot helps you modernize faster. GitHub Copilot's new “upgrade assistant” can help refactor your project, suggest dependency upgrades, and guide you through framework transitions, freeing you up to focus on innovation. Get the Right Fit for Your Java App And what if your apps run on JBoss, WebLogic, or Tomcat? We will walk you through how to map those apps to the right Azure service: Monoliths (JAR, WAR, EAR) → Deploy to App Service Microservices or containers → Use Azure Container Apps or AKS WebLogic & WebSphere → Lift and shift to Azure Virtual Machines JBoss EAP containers → Run on Azure Red Hat OpenShift You’ll get clear guidance on where your apps fit and how to move forward, with no guesswork or dead ends. Let's Code the Future, Together I’ll be there, along with Josh Long from the Spring AI community and Lize Raes from the LangChain4j community, delivering a technical keynote packed with practical insights. If you haven’t started building intelligent Java apps, you can start with JDConf. If you’ve already started on the journey, tune in to learn how you can enrich your experiences with the latest in tech. So, mark your calendar. Spread the word. Bring your team. JDConf 2025 is your place to build what is next with Java and AI. 👉 Register now at jdconf.com. Check out the 20+ exclusive sessions brought to you by Java experts from across the globe in all major time zones.Announcing GA for Azure Container Apps Serverless GPUs
Azure Container Apps Serverless GPUs accelerated by NVIDIA are now generally available. Serverless GPUs enable you to seamlessly run AI workloads with per-second billing and scale down to zero when not in use. Thus, reducing operational overhead to support easy real-time custom model inferencing and other GPU-accelerated workloads. Serverless GPUs accelerate the speed of AI development teams by allowing customers to focus on core AI code and less on managing infrastructure when using GPUs. This provides an excellent middle layer option between Azure AI Model Catalog's serverless APIs and hosting custom models on managed compute. Now customers can build their own serverless API endpoints for inferencing AI models including custom models. Customers can also provision on-demand GPU-powered Jupyter Notebooks or run other compute-intensive AI workloads that are ephemeral in nature. It provides full data governance as customer’s data never leaves the boundaries of the container while still providing a managed, serverless platform from which to build your applications. This GA release of Serverless GPUs also adds support for NVIDIA NIM microservices, NVIDIA NIM™, part of NVIDIA AI Enterprise, is a set of easy-to-use microservices designed for secure, reliable deployment of high-performance AI model inferencing at scale. Supporting a wide range of AI models, including open-source community and NVIDIA AI Foundation models, NVIDIA NIM ensures seamless, scalable AI inferencing leveraging industry-standard APIs. Key benefits of serverless GPUs Scale-to zero GPUs: Support for serverless scaling of NVIDIA A100 and T4 GPUs. Per-second billing: Pay only for the GPU compute you use. Built-in data governance: Your data never leaves the container boundary. Flexible compute options: Choose between NVIDIA A100 and T4 GPUs. Middle-layer for AI development: Bring your own model on a managed, serverless compute platform and easily run your AI applications alongside your existing apps. Scenarios Our customers have been running a wide range of workloads on serverless GPUs. Below are some common use cases. NVIDIA T4 Real-time and batch inferencing: Using custom open-source models with fast startup times, automatic scaling, and a per-second billing model, serverless GPUs are ideal for dynamic applications that don't already have a serverless API in the model catalog. NVIDIA A100 Compute intensive machine learning scenarios: Significantly speed up applications that implement fine-tuned custom generative AI models, deep learning, or neural networks. High performance computing (HPC) and data analytics: Applications that require complex calculations or simulations, such as scientific computing and financial modeling as well as accelerated data processing and analysis among massive datasets. Serverless GPUs with NVIDIA NIM Serverless GPUs now support NVIDIA NIM microservices, which simplify and accelerate the development of AI applications and agentic AI workflows with pre-packaged, scalable, and performance-tuned models that can be deployed as secure inference endpoints on Azure Container Apps. In order to leverage the power of NVIDIA’s NIM, go to NVIDIA’s API catalog: Try NVIDIA NIM APIs, and select the NIM you wish to run with the ‘Run Anywhere’ NIM type. You will need to set your NGC_API_KEY as an environment variable when deploying Azure Container Apps. For a full set of instructions on how to add a NIM to your container app, follow the instructions here. (Note: Each NIM model has certain hardware requirements, Azure Container Apps serverless GPUs support A100 and T4 GPUs. Please ensure the NIM you are selecting is supported by the hardware.) Quota changes for GA With GA, we are introducing default GPU quotas for enterprise and pay-as-you-go customers. All enterprise agreement customers will have quota for A100 and T4 GPUs. The feature is supported in West US 3, Australia East, and Sweden Central. Get started with serverless GPUs From the portal, you can select to enable GPUs for your Consumption app in the container tab when creating your Container App or your Container App Job. Note: In order to achieve the best performance with serverless GPUs, use an Azure Container Registry (ACR) with artifact streaming enabled for your image tag. Follow steps here to enable artifact streaming on your ACR. To learn more about getting started with serverless GPUs, see our quickstart. You can also add a new consumption GPU workload profile to your existing Container App environment through the workload profiles UX in portal or through the CLI commands for managing workload profiles. Learn more about serverless GPUs and NIMs With serverless GPUs, Azure Container Apps now simplifies the development of your AI applications by providing scale-to-zero compute, pay-as you go pricing, reduced infrastructure management, and more. To learn more, visit: Using serverless GPUs in Azure Container Apps (preview) | Microsoft Learn Tutorial: Generate images using serverless GPUs in Azure Container Apps (preview) | Microsoft Learn Tutorial: Deploy an NVIDIA Llama3 NIM to Azure Container Apps Try NVIDIA NIM APIsWhat's New in Azure App Service at Ignite 2024
Learn about the GA of sidecar extensibility on Linux and see team members demonstrating the latest tools for AI assisted web application migration and modernization as well as the latest updates to Java JBoss EAP on Azure App Service. Team members will also demonstrate integrating the Phi-3 small language model with a web application via the new sidecar extensibility using existing App Service hardware! Also new for this year’s Ignite, many topics that attendees see in App Service related sessions are also available for hands-on learning across multiple hands-on labs (HoLs). Don’t just watch team members demonstrating concepts on-stage, drop by one of the many HoL sessions and test drive the functionality yourself! Azure App Service team members will also be in attendance at the Expert Meetup area on the third floor in the Hub – drop by and chat if you are attending in-person! Additional demos, presentations and hands-on labs covering App Service are listed at the end of this blog post for easy reference. Sidecar Extensibility GA for Azure App Service on Linux Sidecar extensibility for Azure App Service on Linux is now GA! Linux applications deployed from source-code as well as applications deployed using custom containers can take advantage of sidecar extensibility. Sidecars enable developers to attach additional capabilities like third-party application monitoring providers, in-memory caches, or even local SLM (small language model) support to their applications without having to bake that functionality directly into their applications. Developers can configure up to four sidecar containers per application, with each sidecar being associated with its own container registry and (optional) startup command. Examples of configuring an OpenTelemetry collector sidecar are available in the documentation for both container-based applications and source-code based applications. There are also several recent blog posts demonstrating additional sidecar scenarios. One example walks through using a Redis cache sidecar as an in-memory cache to accelerate data retrieval in a web application (sample code here). Another example demonstrates adding a sidecar containing the Phi-3 SLM to a custom container web application (sample code here). Once the web app is running with the SLM sidecar, Phi-3 processes text prompts directly on the web server without the need to call remote LLMs or host models on scarce GPU hardware. Similar examples for source deployed applications are available in the Ignite 2024 hands on lab demonstrating sidecars. Exercise three walks through attaching an OTel sidecar to a source-code based application, and exercise four shows how to attach a Phi-3 sidecar to a source-code based application. Looking ahead to the future, App Service will be adding “curated sidecars” to the platform to make it easier for developers to integrate common sidecar scenarios. Development is already underway to include options for popular third-party application monitoring providers, Redis cache support, as well as a curated sidecar encapsulating the Phi-3 SLM example mentioned earlier. Stay tuned for these enhancements in the future! If you are attending Microsoft Ignite 2024 in person, drop by the theater session “Modernize your apps with AI without completely rewriting your code” (session code: THR 614) which demonstrates using sidecar extensibility to add Open Telemetry monitoring as well as Phi-3 SLM support to applications on App Service for Linux! .NET 9 GA, JBoss EAP and More Language Updates! With the recent GA of .NET 9 last week developers can deploy applications running .NET 9 GA on both Windows and Linux variants of App Service! Visual Studio, Visual Studio Code, Azure DevOps and GitHub Actions all support building and deploying .NET 9 applications onto App Service. Start a new project using .NET 9 or upgrade your existing .NET applications in-place and take advantage of .NET 9! For JBoss EAP on App Service for Linux, customers will soon be able to bring their existing JBoss licenses with them when moving JBoss EAP workloads onto App Service for Linux. This change will make it easier and more cost effective than ever for JBoss EAP customers to migrate existing workloads to App Service, including JBoss versions 7.3, 7.4 and 8.0! As a quick reminder, last month App Service also announced reduced pricing for JBoss EAP licenses (for net-new workloads) as well as expanded hardware support (both memory-optimized and Free tier are now supported for JBoss EAP applications). App Service is planning to release both Node 22 and Python 3.13 onto App Service for Linux with expected availability in December! Python 3.13 is the latest stable Python release which means developers will be able to leverage this version with confidence given long term support runs into 2029. Node 22 is the latest active LTS release of Node and is a great version for developers to adopt with its long-term support lasting into 2026. A special note for Linux Python developers, App Service now supports “auto-instrumentation” in public preview for Python versions 3.8 through 3.12. This makes it trivial for source-code based Python applications to enable Application Insights monitoring for their applications by simply turning the feature “on” in the Azure Portal. If you ever thought to yourself that it can be a hassle setting up application monitoring and hence find yourself procrastinating, this is the monitoring feature for you! Looking ahead just a few short weeks until December, App Service also plans to release PHP 8.4 for developers on App Service for Linux. This will enable PHP developers to leverage the latest fully supported PHP release with an expected support cycle stretching into 2028. For WordPress customers Azure App Service has added support for managed identities when connecting to MySQL database as well as storage accounts. The platform has also transitioned WordPress from Alpine Linux to Debian, aligning with App Service for Linux to offer a more secure platform. Looking ahead, App Service is excited to introduce some new features by the end of the year, including an App Service plugin for WordPress! This plugin will enable users to manage WordPress integration with Azure Communication Services email, set up Single Sign-On using Microsoft Entra ID, and diagnose performance bottlenecks. Stay tuned for upcoming WordPress announcements! End-to-End TLS & Min TLS Cipher Suite are now GA End-to-end TLS encryption for public multi-tenant App Service is now GA! When E2E TLS is configured, traffic between the App Service frontends and individual workers is secured using a platform supplied TLS certificate. This additional level of security is available for both Windows and Linux sites using Standard SKU and above as well as Isolatedv2 SKUs. You can enable this feature easily in the Azure Portal by going to your resource, clicking the “Configuration” blade and turning the feature “On” as shown below: Configuration of the minimum TLS cipher suite for a web application is also GA! With this feature developers can choose from a pre-determined list of cipher suites. When a minimum cipher suite is selected, the App Service frontends will reject any incoming requests that use a cipher suite weaker than the selected minimum cipher suite. This feature is supported for both Windows and Linux applications using Basic SKU and higher as well as Isolatedv2 SKUs. You configure a minimum TLS cipher suite in the Azure Portal by going to the “Configuration” blade for a website and selecting “Change” for the Minimum Inbound TLS Cipher Suite setting. In the resulting blade (shown below) you can select the minimum cipher suite for your application: To learn more about these and other TLS features on App Service, please refer to the App Service TLS overview. AI-Powered Conversational Diagnostics Building on the Conversational Diagnostics AI-powered tool and the guided decision making path introduced in Diagnostic Workflows, the team has created a new AI-driven natural language-based diagnostics solution for App Service on Linux. The new solution brings together previous functionality to create an experience that comprehends user intent, selects the appropriate Diagnostic Workflow, and keeps users engaged by providing real-time updates and actionable insights through chat. Conversational Diagnostics also provides the grounding data that the generative AI back-end uses to produce recommendations thus empowering users to check the conclusions. The integration of Conversational Diagnostics and Diagnostic Workflows marks a significant advancement in the platform’s diagnostic capabilities. Stay tuned for more updates and experience the transformative power of Generative AI-driven diagnostics firsthand! App Service Migration and Modernization The team just recently introduced new architectural guidance around evolving and modernizing web applications with the Modern Web Application pattern for .NET and Java! This guidance builds on the Reliable Web App pattern for .NET and Java as well as the Azure Migrate application and code assessment tool. With the newly released Modern Web Application guidance, there is a well-documented path for migrating web applications from on-premises/VM deployments using the application and code assessment tool, iterating and evolving web applications with best practices using guidance from the Reliable Web App pattern, and subsequently going deeper on modernization and re-factoring following guidance from the Modern Web Application pattern. Best of all customers can choose to “enter” this journey at any point and progress as far down the modernization path as needed based on their unique business and technical requirements! As a quick recap on the code assessment tool, it is a guided experience inside of Visual Studio with GitHub Copilot providing actionable guidance and feedback on recommended changes needed to migrate applications to a variety of Azure services including Azure App Service. Combined with AI-powered Conversational Diagnostics (mentioned earlier), developers now have AI-guided journeys supporting them from migration all the way through deployment and runtime operation on App Service! Networking and ASE Updates As of November 1, 2024, we are excited to announce that App Service multi-plan subnet join is generally available across all public Azure regions! Multi-plan subnet join eases network management by reducing subnet sprawl, enabling developers to connect multiple app service plans to a single subnet. There is no limit to the number of app service plans that connect to a single subnet. However, developers should keep in mind the number of available IPs since tasks such as changing the SKU for an app service plan will temporarily double the number of IP addresses used in a connected subnet. For more information as well as examples on using multi-plan subnet join see the documentation! App Service also recently announced GA of memory optimized options for Isolatedv2 on App Service Environment v3. The new memory-optimized options range from two virtual cores with 16 GB RAM in I1mv2 (compared to two virtual cores, 8 GB RAM in I1v2) all the way up to 32 virtual cores with 256 GB RAM in I5mv2. The new plans are available in most regions. Check back regularly to see if your preferred region is supported. For more details on the technical specifications of these plans, as well as information on the complete range of tiers and plans for Microsoft Azure App Service, visit our pricing page. Using services such as Application Gateway and Azure Front Door with App Service as entry points for client traffic is a common scenario that many of our customers implement. However, when using these services together, there are integration challenges around the default cookie domain for HTTP cookies, including the ARRAffinity cookie used for session affinity. App Service collaborated with the Application Gateway team to introduce a simple solution that addresses the session affinity problem. App Service introduced a new session affinity proxy configuration setting in October which tells App Service to always set the hostname for outbound cookies based on the upstream hostname seen by Application Gateway or Azure Front Door. This simplifies integration with a single-click experience for App Service developers who front-end their websites using one of Azure’s reverse proxies, and it solves the challenge of round-tripping the ArrAffinity cookie when upstream proxies are involved. Looking ahead to early 2025, App Service will shortly be expanding support for IPv6 to include both inbound and outbound connections (currently only inbound connections are supported). The current public preview includes dual-stack support for both IPv4 and IPv6, allowing for a smooth transition and compatibility with existing systems. Read more about the latest status of the IPv6 public preview on App Service here ! Lastly, the new application naming and hostname convention that was rolled out a few months earlier for App Service is now GA for App Service. The platform has also extended this new naming convention to Azure Functions where it is now available in public preview for newly created functions. To learn more about the new naming convention and the protection it provides against subdomain takeover take a look at the introductory blog post about the unique default hostname feature. Upcoming Availability Zone Improvements New Availability Zone features are currently rolling out that will make zone redundant App Service deployments more cost efficient and simpler to manage in early 2025! The platform will be changing the minimum requirement for enabling Availability Zones to two instances instead of three, while still maintaining a 99.99% SLA. Many existing app service plans with two or more instances will also automatically become capable of supporting Availability Zones without requiring additional setup. Additionally, the zone redundant setting will be mutable throughout the life of an app service plan. This upcoming improvement will allow customers on Premium V2, Premium V3, or Isolated V2 plans, to toggle zone redundancy on or off as needed. Customers will also gain enhanced visibility into Availability Zone information, including physical zone placement and counts. As a sneak peek into the future, the screenshot below shows what the new experience will look like in the Azure Portal: Stay tuned for Availability Zone updates coming to App Service in early 2025! Next Steps Developers can learn more about Azure App Service at Getting Started with Azure App Service. Stay up to date on new features and innovations on Azure App Service via Azure Updates as well as the Azure App Service (@AzAppService) X feed. There is always a steady stream of great deep-dive technical articles about App Service as well as the breadth of developer focused Azure services over on the Apps on Azure blog. Azure App Service (virtually!) attended the recently completed November .Net Conf 2024. App Service functionality was featured showing a .NET 9.0 app using Azure Sql’s recently released native vector data type support that enables developers to perform hybrid text searches on Azure Sql data using vectors generated via Azure OpenAI embeddings! And lastly take a look at Azure App Service Community Standups hosted on the Microsoft Azure Developers YouTube channel. The Azure App Service Community Standup series regularly features walkthroughs of new and upcoming features from folks that work directly on the product! Ignite 2024 Session Reference (Note: some sessions/labs have more than one timeslot spanning multiple days). (Note: all times below are listed in Chicago time - Central Standard Time). Modernize your apps with AI without completely rewriting your code Modernize your apps with AI without completely rewriting your code [Note: this session includes a demonstration of the Phi-3 sidecar scenario] Wednesday, November 20 th 1:00 PM - 1:30 PM Central Standard Time Theater Session – In-Person Only (THR614) McCormick Place West Building – Level 3, Hub, Theater C Unlock AI: Assess your app and data estate for AI-powered innovation Unlock AI: Assess your app and data estate for AI-powered innovation Wednesday, November 20 th 1:15 PM – 2:00 PM Central Time McCormick Place West Building – Level 1, Room W183c Breakout and Recorded Session (BRK137) Modernize and scale enterprise Java applications on Azure Modernize and scale enterprise Java applications on Azure Thursday, November 21 st 8:30 AM - 9:15 AM Central Time McCormick Place West Building – Level 1, Room W183c Breakout and Recorded Session (BRK147) Assess apps with Azure Migrate and replatform to Azure App Service Assess apps with Azure Migrate and replatform to Azure App Service Tuesday, November 19 th 1:15 PM - 2:30 PM Central Time McCormick Place West Building – Level 4, Room W475 Hands on Lab – In-Person Only (LAB408) Integrate GenAI capabilities into your .NET apps with minimal code changes Integrate GenAI capabilities into your .NET apps with minimal code changes [Note: Lab participants will be able to try out the Phi-3 sidecar scenario in this lab.] Wednesday, November 20 th 8:30 AM - 9:45 AM Central Time McCormick Place West Building – Level 4, Room W475 Hands on Lab – In-Person Only (LAB411) Assess apps with Azure Migrate and replatform to Azure App Service Assess apps with Azure Migrate and replatform to Azure App Service Wednesday, November 20 th 6:30 PM - 7:45 PM Central Time McCormick Place West Building – Level 4, Room W470b Hands on Lab – In-Person Only (LAB408-R1) Integrate GenAI capabilities into your .NET apps with minimal code changes Integrate GenAI capabilities into your .NET apps with minimal code changes [Note: Lab participants will be able to try out the Phi-3 sidecar scenario in this lab.] Thursday, November 21 st 10:15 AM - 11:30 AM Central Time McCormick Place West Building – Level 1, Room W180 Hands on Lab – In-Person Only (LAB411-R1) Assess apps with Azure Migrate and replatform to Azure App Service Assess apps with Azure Migrate and replatform to Azure App Service Friday, November 22 nd 9:00 AM – 10:15 AM Central Time McCormick Place West Building – Level 4, Room W474 Hands on Lab – In-Person Only (LAB408-R2)
