Reimagining AI Ops with Azure SRE Agent: New Automation, Integration, and Extensibility features
Azure SRE Agent offers intelligent and context aware automation for IT operations. Enhanced by customer feedback from our preview, the SRE Agent has evolved into an extensible platform to automate and manage tasks across Azure and other environments. Built on an Agentic DevOps approach - drawing from proven practices in internal Azure operations - the Azure SRE Agent has already saved over 20,000 engineering hours across Microsoft product teams operations, delivering strong ROI for teams seeking sustainable AIOps. An Operations Agent that adapts to your playbooks Azure SRE Agent is an AI powered operations automation platform that empowers SREs, DevOps, IT operations, and support teams to automate tasks such as incident response, customer support, and developer operations from a single, extensible agent. Its value proposition and capabilities have evolved beyond diagnosis and mitigation of Azure issues, to automating operational workflows and seamless integration with the standards and processes used in your organization. SRE Agent is designed to automate operational work and reduce toil, enabling developers and operators to focus on high-value tasks. By streamlining repetitive and complex processes, SRE Agent accelerates innovation and improves reliability across cloud and hybrid environments. In this article, we will look at what’s new and what has changed since the last update. What’s New: Automation, Integration, and Extensibility Azure SRE Agent just got a major upgrade. From no-code automation to seamless integrations and expanded data connectivity, here’s what’s new in this release: No-code Sub-Agent Builder: Rapidly create custom automations without writing code. Flexible, event-driven triggers: Instantly respond to incidents and operational changes. Expanded data connectivity: Unify diagnostics and troubleshooting across more data sources. Custom actions: Integrate with your existing tools and orchestrate end-to-end workflows via MCP. Prebuilt operational scenarios: Accelerate deployment and improve reliability out of the box. Unlike generic agent platforms, Azure SRE Agent comes with deep integrations, prebuilt tools, and frameworks specifically for IT, DevOps, and SRE workflows. This means you can automate complex operational tasks faster and more reliably, tailored to your organization’s needs. Sub-Agent Builder: Custom Automation, No Code Required Empower teams to automate repetitive operational tasks without coding expertise, dramatically reducing manual workload and development cycles. This feature helps address the need for targeted automation, letting teams solve specific operational pain points without relying on one-size-fits-all solutions. Modular Sub-Agents: Easily create custom sub-agents tailored to your team’s needs. Each sub-agent can have its own instructions, triggers, and toolsets, letting you automate everything from outage response to customer email triage. Prebuilt System Tools: Eliminate the inefficiency of creating basic automation from scratch, and choose from a rich library of hundreds of built-in tools for Azure operations, code analysis, deployment management, diagnostics, and more. Custom Logic: Align automation to your unique business processes by defining your automation logic and prompts, teaching the agent to act exactly as your workflow requires. Flexible Triggers: Automate on Your Terms Invoke the agent to respond automatically to mission-critical events, not wait for manual commands. This feature helps speed up incident response and eliminate missed opportunities for efficiency. Multi-Source Triggers: Go beyond chat-based interactions, and trigger the agent to automatically respond to Incident Management and Ticketing systems like PagerDuty and ServiceNow, Observability Alerting systems like Azure Monitor Alerts, or even on a cron-based schedule for proactive monitoring and best-practices checks. Additional trigger sources such as GitHub issues, Azure DevOps pipelines, email, etc. will be added over time. This means automation can start exactly when and where you need it. Event-Driven Operations: Integrate with your CI/CD, monitoring, or support systems to launch automations in response to real-world events - like deployments, incidents, or customer requests. Vital for reducing downtime, it ensures that business-critical actions happen automatically and promptly. Expanded Data Connectivity: Unified Observability and Troubleshooting Integrate data, enabling comprehensive diagnostics and troubleshooting and faster, more informed decision-making by eliminating silos and speeding up issue resolution. Multiple Data Sources: The agent can now read data from Azure Monitor, Log Analytics, and Application Insights based on its Azure role-based access control (RBAC). Additional observability data sources such as Dynatrace, New Relic, Datadog, and more can be added via the Remote Model Context Protocol (MCP) servers for these tools. This gives you a unified view for diagnostics and automation. Knowledge Integration: Rather than manually detailing every instruction in your prompt, you can upload your Troubleshooting Guide (TSG) or Runbook directly, allowing the agent to automatically create an execution plan from the file. You may also connect the agent to resources like SharePoint, Jira, or documentation repositories through Remote MCP servers, enabling it to retrieve needed files on its own. This approach utilizes your organization’s existing knowledge base, streamlining onboarding and enhancing consistency in managing incidents. Azure SRE Agent is also building multi-agent collaboration by integrating with PagerDuty and Neubird, enabling advanced, cross-platform incident management and reliability across diverse environments. Custom Actions: Automate Anything, Anywhere Extend automation beyond Azure and integrate with any tool or workflow, solving the problem of limited automation scope and enabling end-to-end process orchestration. Out-of-the-Box Actions: Instantly automate common tasks like running azcli, kubectl, creating GitHub issues, or updating Azure resources, reducing setup time and operational overhead. Communication Notifications: The SRE Agent now features built-in connectors for Outlook, enabling automated email notifications, and for Microsoft Teams, allowing it to post messages directly to Teams channels for streamlined communication. Bring Your Own Actions: Drop in your own Remote MCP servers to extend the agent’s capabilities to any custom tool or workflow. Future-proof your agentic DevOps by automating proprietary or emerging processes with confidence. Prebuilt Operations Scenarios Address common operational challenges out of the box, saving teams time and effort while improving reliability and customer satisfaction. Incident Response: Minimize business impact and reduce operational risk by automating detection, diagnosis, and mitigation of your workload stack. The agent has built-in runbooks for common issues related to many Azure resource types including Azure Kubernetes Service (AKS), Azure Container Apps (ACA), Azure App Service, Azure Logic Apps, Azure Database for PostgreSQL, Azure CosmosDB, Azure VMs, etc. Support for additional resource types is being added continually, please see product documentation for the latest information. Root Cause Analysis & IaC Drift Detection: Instantly pinpoint incident causes with AI-driven root cause analysis including automated source code scanning via GitHub and Azure DevOps integration. Proactively detect and resolve infrastructure drift by comparing live cloud environments against source-controlled IaC, ensuring configuration consistency and compliance. Handle Complex Investigations: Enable the deep investigation mode that uses a hypothesis-driven method to analyze possible root causes. It collects logs and metrics, tests hypotheses with iterative checks, and documents findings. The process delivers a clear summary and actionable steps to help teams accurately resolve critical issues. Incident Analysis: The integrated dashboard offers a comprehensive overview of all incidents managed by the SRE Agent. It presents essential metrics, including the number of incidents reviewed, assisted, and mitigated by the agent, as well as those awaiting human intervention. Users can leverage aggregated visualizations and AI-generated root cause analyses to gain insights into incident processing, identify trends, enhance response strategies, and detect areas for improvement in incident management. Inbuilt Agent Memory: The new SRE Agent Memory System transforms incident response by institutionalizing the expertise of top SREs - capturing, indexing, and reusing critical knowledge from past incidents, investigations, and user guidance. Benefit from faster, more accurate troubleshooting, as the agent learns from both successes and mistakes, surfacing relevant insights, runbooks, and mitigation strategies exactly when needed. This system leverages advanced retrieval techniques and a domain-aware schema to ensure every on-call engagement is smarter than the last, reducing mean time to resolution (MTTR) and minimizing repeated toil. Automatically gain a continuously improving agent that remembers what works, avoids past pitfalls, and delivers actionable guidance tailored to the environment. GitHub Copilot and Azure DevOps Integration: Automatically triage, respond to, and resolve issues raised in GitHub or Azure DevOps. Integration with modern development platforms such as GitHub Copilot coding agent increases efficiency and ensures that issues are resolved faster, reducing bottlenecks in the development lifecycle. Ready to get started? Azure SRE Agent home page Product overview Pricing Page Pricing Calculator Pricing Blog Demo recordings Deployment samples What’s Next? Give us feedback: Your feedback is critical - You can Thumbs Up / Thumbs Down each interaction or thread, or go to the “Give Feedback” button in the agent to give us in-product feedback - or you can create issues or just share your thoughts in our GitHub repo at https://github.com/microsoft/sre-agent. We’re just getting started. In the coming months, expect even more prebuilt integrations, expanded data sources, and new automation scenarios. We anticipate continuous growth and improvement throughout our agentic AI platforms and services to effectively address customer needs and preferences. Let us know what Ops toil you want to automate next!Azure SRE Agent: Expanding Observability and Multi-Cloud Resilience
The Azure SRE Agent continues to evolve as a cornerstone for operational excellence and incident management. Over the past few months, we have made significant strides in enabling integrations with leading observability platforms—Dynatrace, New Relic, and Datadog—through Model Context Protocol (MCP) Servers. These partnerships serve joint customers, enabling automated remediation across diverse environments. Deepening Integrations with MCP Servers Our collaboration with these partners is more than technical—it’s about delivering value at scale. Datadog, New Relic, and Dynatrace are all Azure Native ISV Service partners. With these integrations Azure Native customers can also choose to add these MCP servers directly from the Azure Native partners’ resource: Datadog: At Ignite, Azure SRE Agent was presented with the Datadog MCP Server, to demonstrate how our customers can streamline complex workflows. Customers can now bring their Datadog MCP Server into Azure SRE Agent, extending knowledge capabilities and centralizing logs and metrics. Find Datadog Azure Native offerings on Marketplace. New Relic: When an alert fires in New Relic, the Azure SRE Agent calls the New Relic MCP Server to provide Intelligent Observability insights. This agentic integration with the New Relic MCP Server offers over 35 specialized tools across, entity and account management, alerts and monitoring, data analysis and queries, performance analysis, and much more. The advanced remediation skills of the Azure SRE Agent + New Relic AI help our joint customers diagnose and resolve production issues faster. Find New Relic’s Azure Native offering on Marketplace Dynatrace: The Dynatrace integration bridges Microsoft Azure's cloud-native infrastructure management with Dynatrace's AI-powered observability platform, leveraging the Davis AI engine and remote MCP server capabilities for incident detection, root cause analysis, and remediation across hybrid cloud environments. Check out Dynatrace’s Azure Native offering on Marketplace. These integrations are made possible by Azure SRE Agent’s MCP connectors. The MCP connectors in Azure SRE Agent act as the bridge between the agent and MCP servers, enabling dynamic discovery and execution of specialized tools for observability and incident management across diverse environments. This feature allows customers to build their own custom sub-agents to leverage tools from MCP Servers from integrated platforms like Dynatrace, Datadog, and New Relic to complement the agent’s diagnostic and remediation capabilities. By connecting Azure SRE Agent to external MCP servers scenarios such as cross-platform telemetry analysis are unlocked. Looking Ahead: Multi-Agent Collaboration Azure SRE Agent isn’t stopping with MCP integrations We’re actively working with PagerDuty and NeuBird to support dynamic use cases via agent-to-agent collaboration: PagerDuty: PagerDuty’s PD Advance SRE Agent is an AI-powered assistant that triages incidents by analyzing logs, diagnostics, past incident history, and runbooks to surface relevant context and recommended remediations. At Ignite, PagerDuty and Microsoft demonstrated how Azure SRE Agent can ingest PagerDuty incidents and collaborate with PagerDuty’s SRE Agent to complement triage using historical patterns, runbook intelligence and Azure diagnostics. NeuBird: NeuBird’s Agentic AI SRE, Hawkeye, autonomously investigates and resolves incidents across hybrid, and multi-cloud environments. By connecting to telemetry sources like Azure Monitor, Prometheus, and GitHub, Hawkeye delivers real-time diagnosis and targeted fixes. Building on the work presented at SRE Day this partnership underscores our commitment to agentic ecosystems where specialized agents collaborate for complex scenarios. Sign up for the private preview to try the integration, here. Additionally, please check out NeuBird on Marketplace. These efforts reflect a broader vision: Azure SRE Agent as a hub for cross-platform reliability, enabling customers to manage incidents across Azure, on-premises, and other clouds with confidence. Why This Matters As organizations embrace distributed architectures, the need for integrated, intelligent, and multi-cloud-ready SRE solutions has never been greater. By partnering with industry leaders and pioneering agent-to-agent workflows, Azure SRE Agent is setting the stage for a future where resilience is not just reactive—it’s proactive and collaborative.The City Leader's Dilemma: How AI Is turning urban strain into strategic advantage
Ready to transform how your city plans and operates? Download the Trend Report 2025: Planning and operating thriving cities – innovation for smarter urban living to access the complete playbook on AI-powered urban innovation, complete with case studies from Bangkok, Singapore, Barcelona, and Manchester. Urban challenges aren’t slowing down. Populations are growing, climate pressures are intensifying, and residents expect seamless services, while budgets remain flat and workforces stretch thin. Traditional approaches can’t keep pace. The good news? Cities worldwide are showing that AI and digital innovation can drive meaningful improvements. Recent studies indicate that more than half of surveyed cities are already using AI to upgrade operations, and most plan to expand adoption in the next three years. For many leaders, the question is less about whether to act and more about how to act responsibly and effectively. After studying the latest research and real-world deployments, three strategic shifts stand out, each offering a different lens on how forward-thinking city leaders are turning pressure into progress. Shift One: From Fragmented services to unified citizen experiences Residents expect seamless problem-solving, not organizational complexity. Yet many cities operate in silos, transit systems, permitting offices, 311 reporting, and community engagement often run on separate platforms. The result? Multiple apps for residents, duplicated effort for staff, and missed insights locked in departmental databases. Leading cities are breaking this pattern through unified digital platforms powered by AI. Bangkok’s Traffy Fondue: Citizens report issues like broken streetlights or flooding via a mobile interface. AI categorizes each report and routes it to the right department. By mid-2025, the platform handled nearly one million citizen reports, improving engagement and reducing administrative overhead. The outcome? Reduced administrative overhead, and something harder to measure but equally important: residents who believe their government actually listens. Buenos Aires took a similar path with "Boti," a WhatsApp chatbot that evolved from a COVID-era tool into a citywide digital assistant. Citizens report issues, ask questions, and access services through the messaging app they already use daily. Technology that meets residents where they are improves efficiency and strengthens trust, when guided by principles of transparency and fairness. Shift Two: From reactive planning to predictive foresight Traditional urban planning relies on static models: masterplans, zoning maps, historical growth trends. These tools served their purpose. But they cannot capture the complexity of future risks, extreme weather, evolving mobility patterns, or the cascading effects of a single development decision. Digital twins complement human expertise by integrating geospatial data, climate models, and policy scenarios, helping cities make smarter decisions with limited budgets. Singapore's Digital Urban Climate Twin integrates geospatial data with climate models to simulate how different policies would affect temperature and thermal comfort across neighborhoods. These tools support informed decision-making while maintaining human oversight and accountability. The result? Strategic adaptation rather than reactive firefighting. Sydney built an urban digital twin that correlates environmental conditions with traffic accidents, using machine learning to predict crash risk on specific road segments. City planners can now test interventions virtually, what happens if we lower speed limits here? Add a bike lane there? Before committing resources. Even smaller cities are finding value. Imola, Italy uses a microclimate digital twin to model heat distribution street by street, guiding decisions about where to plant trees or specify cool pavement materials. The paradigm shift is profound: instead of planning based on what happened, cities can now plan based on what's likely to happen. This is how you make smart bets with limited budgets. Shift Three: From tech adoption to governance architecture Here's where many cities stumble. They invest in flashy pilots without building the institutional structures to sustain them. The cities getting this right treat governance as a strategic asset, not a compliance burden. Singapore's Model AI Governance Framework provides practical guidelines for transparency, fairness, and human-centric design. Its AI Verify toolkit lets organizations test their systems for resilience, accountability, and bias before deployment. Barcelona takes a different but equally rigorous approach, treating municipal data as a public asset under its Data Commons program. The city's procurement strategy favors open-source solutions, preventing vendor lock-in while supporting local innovation ecosystems. Both models share a common insight: rapid innovation doesn't automatically produce equitable outcomes. Governance creates the guardrails that allow experimentation without derailment. For city leaders, this means building cross-sector governance councils, adopting clear data strategies, creating ethical AI frameworks, and investing in workforce capability. These aren't obstacles to innovation; they're the foundation that makes sustained innovation possible. The Path Forward Cities that thrive in combine strategic vision with disciplined, responsible technology use. They embed digital capabilities into decision-making, supported by robust policies and cross-department collaboration. Learn how Microsoft helps governments build tech-empowered cities and resilient infrastructure at Microsoft for government. The Smart Cities World 2025 Trend Report provides the detailed case studies, governance frameworks, and implementation roadmaps to make this real. Download your copy now and start building the city your residents deserve.
Proactive Monitoring Made Simple with Azure SRE Agent
SRE teams strive for proactive operations, catching issues before they impact customers and reducing the chaos of incident response. While perfection may be elusive, the real goal is minimizing outages and gaining immediate line of sight into production environments. Today, that’s harder than ever. It requires correlating countless signals and alerts, understanding how they relate—or don’t relate—to each other, and assigning the right sense of urgency and impact. Anything that shortens this cycle, accelerates detection, and enables automated remediation is what modern SRE teams crave. What if you could skip the scripting and pipelines? What if you could simply describe what you want in plain language and let it run automatically on a schedule? Scheduled Tasks for Azure SRE Agent With Scheduled Tasks for Azure SRE Agent, that what-if scenario is now a reality. Scheduled tasks combine natural language prompts with Azure SRE Agent’s automation capabilities, so you can express intent, set a schedule, and let the agent do the rest—without writing a single line of code. This means: ⚡ Faster incident response through early detection ✅ Better compliance with automated checks 🎯 More time for high-value engineering work and innovation 💡 The shift from reactive to proactive: Instead of waiting for alerts to fire or customers to report issues, you’re continuously monitoring, validating, and catching problems before they escalate. How Scheduled Tasks Work Under the Hood When you create a Scheduled Task, the process is more than just running a prompt on a timer. Here’s what happens: 1. Prompt Interpretation and Plan Creation The SRE Agent takes your natural language prompt—such as “Scan all resources for security best practices”—and converts it into a structured execution plan. This plan defines the steps, tools, and data sources required to fulfill your request. 2. Built-In Tools and MCP Integration The agent uses its built-in capabilities (Azure CLI, Log Analytics workspace, Appinsights) and can also leverage 3 rd party data sources or tools via MCP server integration for extended functionality. 3. Results Analysis and Smart Summarization After execution, the agent analyzes results, identifies anomalies or issues, and provides actionable summaries not just raw data dumps. 4. Notification and Escalation Based on findings, the agent can: Post updates to Teams channels Create or update incidents Send email notifications Trigger follow-up actions Real-World Use Cases for Proactive Ops Here’s where scheduled tasks shine for SRE teams: Use Case Prompt Example Schedule Security Posture Check “Scan all subscriptions for resources with public endpoints and flag any that shouldn’t be exposed” Daily Cost Anomaly Detection “Compare this week’s spend against last week and alert if any service exceeds 20% growth” Weekly Compliance Drift Detection “Check all storage accounts for encryption settings and report any non-compliant resources” Daily Resource Health Summary “Summarize the health status of all production VMs and highlight any degraded instances” Every 4 hours Incident Trend Analysis “Analyze ICM incidents from the past week, identify patterns, and summarize top contributing services” Weekly Getting Started in 3 Steps Step 1: Define Your Intent Write a natural language prompt describing what you want to monitor or check. Be specific about: - What resources or scope - What conditions to look for - What action to take if issues are found Example: > “Every morning at 8 AM, check all production Kubernetes clusters for pods in CrashLoopBackOff state. If any are found, post a summary to the #sre-alerts Teams channel with cluster name, namespace, and pod details.” Step 2: Set Your Schedule Choose how often the task should run: - Cron expressions for precise control - Simple intervals (hourly, daily, weekly) Step 3: Define Where to Receive Updates Include in your prompt where you want results delivered when the task finishes execution. The agent can use its built-in tools and connectors to: - Post summaries to a Teams channel - Send email notifications - Create or update ICM incidents Example prompt with notification: > "Check all production databases for long-running queries over 30 seconds. If any are found, post a summary to the #database-alerts Teams channel." Why This Matters for Proactive Operations Traditional monitoring approaches have limitations: Traditional Approach With Scheduled Tasks Write scripts, maintain pipelines Describe in plain language Static thresholds and rules Contextual, AI-powered analysis Alert fatigue from noisy signals Smart summarization of what matters Separate tools for check vs. action Unified detection and response Requires dedicated DevOps effort Any SRE can create and modify The result? Your team spends less time building and maintaining monitoring infrastructure and more time on the work that truly requires human expertise. Best Practices for Scheduled Tasks Start simple, iterate — Begin with one or two high-value checks and expand as you gain confidence Be specific in prompts — The more context you provide, the better the results Set appropriate frequencies — Not everything needs to run hourly; match the schedule to the risk Review and refine — Check task results periodically and adjust prompts for better accuracy What’s Next? Scheduled tasks are just the beginning. We’re continuing to invest in capabilities that help SRE teams shift left—catching issues earlier, automating routine checks, and freeing up time for strategic work. Ready to Start? Use this sample that shows how to create a scheduled health check sub-agent: https://github.com/microsoft/sre-agent/blob/main/samples/automation/samples/02-scheduled-health-check-sample.md This example demonstrates: - Building a HealthCheckAgent using built-in tools like Azure CLI and Log Analytics Workspace - Scheduling daily health checks for a container app at 7 AM - Sending email alerts when anomalies are detected 🔗 Explore more samples here: https://github.com/microsoft/sre-agent/tree/main/samples More to Learn Ignite 2025 announcements: https://aka.ms/ignite25/blog/sreagent Documentation: https://aka.ms/sreagent/docs Support & Feature Requests: https://github.com/microsoft/sre-agent/issuesDeploying a Bun + Hono + Vite app to Azure App Service
TOC Introduction Local Environment Deployment Conclusion 1. Introduction Anthropic, the company behind Claude, recently acquired the JavaScript runtime startup Bun, marking one of the most significant shifts in the modern JavaScript ecosystem since the arrival of Node.js and Deno. This acquisition signals more than a business move, it represents a strategic consolidation of performance-oriented tooling, developer ergonomics, and the future of AI-accelerated software development. At the center of this momentum lies a powerful trio: Bun, Hono, and Vite. Bun is a next-generation JavaScript runtime that reimagines everything from dependency installation to HTTP servers, bundling, and execution speed. On top of Bun, frameworks like Hono provide an elegant, lightweight approach to building APIs and full web applications. Hono embraces the Web Standard API model while optimizing for speed and minimal footprint. For front-end development, Vite completes the trio. Vite provides lightning-fast local development through native ES modules and an optimized build pipeline. When paired with Bun, the developer experience becomes even smoother, as Bun accelerates not only the dev server but also the entire build process. The result is a full-stack workflow where front-end and back-end both benefit from a consistent, high-performance environment. This article will guide you through deploying a Bun + Hono + Vite application to an Azure Linux Web App. 2. Local Environment The development environment used in this example is a Docker container. All project creation, modification, and testing will take place inside this environment. We will build an application using Bun + Hono + Vite, containing three endpoints: / : The root endpoint, which displays Hello Bun Hono Vite /api/hello : A static endpoint /api/backend : A runtime endpoint executed on the server, returning computed results for the frontend to display Create the Docker development environment This command generates a Bun + Hono + Vite project. docker run --rm -it -v "$PWD":/app -w /app oven/bun:latest bunx create-vite . --template vanilla-ts Follow the prompts as shown in the image and make the appropriate selections. After completing the prompts, the corresponding project files will be created. At this point, you may press Ctrl + C to stop the running Bun server. Next, install Hono: docker run --rm -it -v "$PWD":/app -w /app oven/bun:latest bun add hono Create 4 files and edit 2 files Create .vscode/settings.json This file serves two purposes: To prevent the large node_modules folder from being uploaded during deployment. Uploading it wastes bandwidth and may cause compatibility issues between local and production environments. To disable ORYX BUILD from interfering in the deployment process. We will use a custom startup script to handle all build tasks instead. { "appService.zipIgnorePattern": [ "node_modules{,/**}", ".git{,/**}", ".vscode{,/**}" ], "appService.showBuildDuringDeployPrompt": false } Create vite.config.ts This file configures Vite’s base path to use relative paths instead of absolute paths. Although this does not affect the production environment, it is essential locally when URLs and ports may differ. import { defineConfig } from 'vite'; export default defineConfig({ base: './', }); Create server.ts This file configures backend routing using Hono and sets up the Bun server. It includes several test endpoints, some static, and some executed at runtime. import { Hono } from 'hono'; import { serveStatic } from 'hono/bun'; const app = new Hono(); app.get('/api/hello', (c) => { return c.text('this is api/hello'); }); app.get('/api/backend', (c) => { const result = 1 + 1; return c.json({ message: "this is /api/backend", calc: `1 + 1 = ${result}`, value: result, }); }); app.use('/assets/*', serveStatic({ root: './dist' })); app.use('/*', serveStatic({ root: './dist' })); app.get('/', serveStatic({ path: './dist/index.html' })); const port = Number(process.env.PORT ?? 3000); export default { port, fetch: app.fetch, }; Create startup.sh This script serves several important roles: Remove any node_modules folders or tar archives created by ORYX during deployment Install the Bun runtime Fully take over the ORYX build process Start the Bun server as the final step #!/bin/bash set -e echo "===== Startup script running =====" cd /home/site/wwwroot echo "Cleaning up Oryx-generated node_modules..." if [ -d /node_modules ]; then echo "Removing /node_modules ..." rm -rf /node_modules fi if [ -f node_modules.tar.gz ]; then echo "Removing node_modules.tar.gz ..." rm -f node_modules.tar.gz fi echo "Oryx cleanup complete." export BUN_INSTALL=/home/site/wwwroot/.bun export PATH="$BUN_INSTALL/bin:/home/site/wwwroot/node_modules/.bin:$PATH" export NODE_PATH="/home/site/wwwroot/node_modules" if [ ! -f "$BUN_INSTALL/bin/bun" ]; then echo "Bun not found. Installing..." curl -fsSL https://bun.sh/install | bash else echo "Bun already installed at $BUN_INSTALL" fi echo "Using Bun version:" bun --version echo "Running bun install ..." bun install echo "Running bun run build ..." bun run build echo "Starting server with bun run start ..." bun run start Modify src/main.ts Simplify the default welcome page so it only displays our test text. // src/main.ts document.querySelector<HTMLDivElement>('#app')!.innerHTML = ` <h1>Hello Bun Hono Vite</h1> `; Modify package.json Update the scripts section so that the project uses the newly created server.ts as the server entry point. { "name": "app", "private": true, "version": "0.0.0", "type": "module", "scripts": { "dev:vite": "vite", "build": "vite build", "preview": "vite preview", "start": "bun server.ts" }, "devDependencies": { "typescript": "~5.9.3", "vite": "^7.2.4" }, "dependencies": { "hono": "^4.10.7" } } Build the project locally This command generates a dist directory containing all Vite-built static assets. docker run --rm -it -v "$PWD":/app -w /app oven/bun:latest bun run build Run the server locally for testing docker run --rm -it -v "$PWD":/app -w /app -p 3000:3000 -e PORT=3000 oven/bun:latest bun run start You can press Ctrl + C to stop the server when finished. http://127.0.0.1:3000/ http://127.0.0.1:3000/api/hello http://127.0.0.1:3000/api/backend 3. Deployment We create a Linux Web App with a minimum SKU of Premium. Add Environment Variables SCM_DO_BUILD_DURING_DEPLOYMENT=false Purpose: Prevents the deployment environment from packaging during publish. This must also be set in the deployment environment itself. WEBSITE_RUN_FROM_PACKAGE=false Purpose: Instructs Azure Web App not to run the app from a prepackaged file. ENABLE_ORYX_BUILD=false Purpose: Prevents Azure Web App from building after deployment. All build tasks will instead execute during the startup script. Add Startup Command bash /home/site/wwwroot/startup.sh After that, we can return to VS Code and deploy the project. Once the deployment is complete, wait about five minutes for the build process to finish, and then you can begin testing. / /api/hello /api/backend 4. Conclusion Bun + Hono + Vite form a cohesive ecosystem that embodies the next era of JavaScript development: fast, compact, ergonomic, and deeply aligned with modern infrastructure needs. It is particularly well-suited for AI applications, where latency, concurrency, and rapid iteration matter more than ever. From streaming inference endpoints to vector database integrations, this stack offers the responsiveness and scalability essential for AI-powered systems.Common Misconceptions When Running Locally vs. Deploying to Azure Linux-based Web Apps
TOC Introduction Environment Variable Build Time Compatible Memory Conclusion 1. Introduction One of the most common issues during project development is the scenario where “the application runs perfectly in the local environment but fails after being deployed to Azure.” In most cases, deployment logs will clearly reveal the problem and allow you to fix it quickly. However, there are also more complicated situations where "due to the nature of the error itself" relevant logs may be difficult to locate. This article introduces several common categories of such problems and explains how to troubleshoot them. We will demonstrate them using Python and popular AI-related packages, as these tend to exhibit compatibility-related behavior. Before you begin, it is recommended that you read Deployment and Build from Azure Linux based Web App | Microsoft Community Hub on how Azure Linux-based Web Apps perform deployments so you have a basic understanding of the build process. 2. Environment Variable Simulating a Local Flask + sklearn Project First, let’s simulate a minimal Flask + sklearn project in any local environment (VS Code in this example). For simplicity, the sample code does not actually use any sklearn functions; it only displays plain text. app.py from flask import Flask app = Flask(__name__) @app.route("/") def index(): return "hello deploy environment variable" if __name__ == "__main__": app.run(host="0.0.0.0", port=8000) We also preset the environment variables required during Azure deployment, although these will not be used when running locally. .deployment [config] SCM_DO_BUILD_DURING_DEPLOYMENT=false As you may know, the old package name sklearn has long been deprecated in favor of scikit-learn. However, for the purpose of simulating a compatibility error, we will intentionally specify the outdated package name. requirements.txt Flask==3.1.0 gunicorn==23.0.0 sklearn After running the project locally, you can open a browser and navigate to the target URL to verify the result. python3 -m venv .venv source .venv/bin/activate pip install -r requirements.txt python app.py Of course, you may encounter the same compatibility issue even in your local environment. Simply running the following command resolves it: export SKLEARN_ALLOW_DEPRECATED_SKLEARN_PACKAGE_INSTALL=True We will revisit this error and its solution shortly. For now, create a Linux Web App running Python 3.12 and configure the following environment variables. We will define Oryx Build as the deployment method. SCM_DO_BUILD_DURING_DEPLOYMENT=false WEBSITE_RUN_FROM_PACKAGE=false ENABLE_ORYX_BUILD=true After deploying the code and checking the Deployment Center, you should see an error similar to the following. From the detailed error message, the cause is clear: sklearn is deprecated and replaced by scikit-learn, so additional compatibility handling is now required by the Python runtime. The error message suggests the following solutions: Install the newer scikit-learn package directly. If your project is deeply coupled to the old sklearn package and cannot be refactored yet, enable compatibility by setting an environment variable to allow installation of the deprecated package. Typically, this type of “works locally but fails on Azure” behavior happens because the deprecated package was installed in the local environment a long time ago at the start of the project, and everything has been running smoothly since. Package compatibility issues like this are very common across various languages on Linux. When a project becomes tightly coupled to an outdated package, you may not be able to upgrade it immediately. In these cases, compatibility workarounds are often the only practical short-term solution. In our example, we will add the environment variable: SKLEARN_ALLOW_DEPRECATED_SKLEARN_PACKAGE_INSTALL=True However, here comes the real problem: This variable is needed during the build phase, but the environment variables set in Azure Portal’s Application Settings only take effect at runtime. So what should we do? The answer is simple, shift the Oryx Build process from build-time to runtime. First, open Azure Portal → Configuration and disable Oryx Build. ENABLE_ORYX_BUILD=false Next, modify the project by adding a startup script. run.sh #!/bin/bash export SKLEARN_ALLOW_DEPRECATED_SKLEARN_PACKAGE_INSTALL=True python -m venv .venv source .venv/bin/activate pip install -r requirements.txt python app.py The startup script works just like the commands you run locally before executing the application. The difference is that you can inject the necessary compatibility environment variables before running pip install or starting the app. After that, return to Azure Portal and add the following Startup Command under Stack Settings. This ensures that your compatibility environment variables and build steps run before the runtime starts. bash run.sh Your overall project structure will now look like this. Once redeployed, everything should work correctly. 3. Build Time Build-Time Errors Caused by AI-Related Packages Many build-time failures are caused by AI-related packages, whose installation processes can be extremely time-consuming. You can investigate these issues by reviewing the deployment logs at the following maintenance URL: https://<YOUR_APP_NAME>.scm.azurewebsites.net/newui Compatible Let’s simulate a Flask + numpy project. The code is shown below. app.py from flask import Flask app = Flask(__name__) @app.route("/") def index(): return "hello deploy compatible" if __name__ == "__main__": app.run(host="0.0.0.0", port=8000) We reuse the same environment variables from the sklearn example. .deployment [config] SCM_DO_BUILD_DURING_DEPLOYMENT=false This time, we simulate the incompatibility between numpy==1.21.0 and Python 3.10. requirements.txt Flask==3.1.0 gunicorn==23.0.0 numpy==1.21.0 We will skip the local execution part and move directly to creating a Linux Web App running Python 3.10. Configure the same environment variables as before, and define the deployment method as runtime build. SCM_DO_BUILD_DURING_DEPLOYMENT=false WEBSITE_RUN_FROM_PACKAGE=false ENABLE_ORYX_BUILD=false After deployment, Deployment Center shows a successful publish. However, the actual website displays an error. At this point, you must check the deployment log files mentioned earlier. You will find two key logs: 1. docker.log Displays real-time logs of the platform creating and starting the container. In this case, you will see that the health probe exceeded the default 230-second startup window, causing container startup failure. This tells us the root cause is container startup timeout. To determine why it timed out, we must inspect the second file. 2. default_docker.log Contains the internal execution logs of the container. Not generated in real time, usually delayed around 15 minutes. Therefore, if docker.log shows a timeout error, wait at least 15 minutes to allow the logs to be written here. In this example, the internal log shows that numpy was being compiled during pip install, and the compilation step took too long. We now have a concrete diagnosis: numpy 1.21.0 is not compatible with Python 3.10, which forces pip to compile from source. The compilation exceeds the platform’s startup time limit (230 seconds) and causes the container to fail. We can verify this by checking numpy’s official site: numpy · PyPI numpy 1.21.0 only provides wheels for cp37, cp38, cp39 but not cp310 (which is python 3.10). Thus, compilation becomes unavoidable. Possible Solutions Set the environment variable WEBSITES_CONTAINER_START_TIME_LIMIT to increase the allowed container startup time. Downgrade Python to 3.9 or earlier. Upgrade numpy to 1.21.0+, where suitable wheels for Python 3.10 are available. In this example, we choose this option. After upgrading numpy to version 1.25.0 (which supports Python 3.10) from specifying in requirements.txt and redeploying, the issue is resolved. numpy · PyPI requirements.txt Flask==3.1.0 gunicorn==23.0.0 numpy==1.25.0 Memory The final example concerns the App Service SKU. AI packages such as Streamlit, PyTorch, and others require significant memory. Any one of these packages may cause the build process to fail due to insufficient memory. The error messages vary widely each time. If you repeatedly encounter unexplained build failures, check Deployment Center or default_docker.log for Exit Code 137, which indicates that the system ran out of memory during the build. The only solution in such cases is to scale up. 4. Conclusion This article introduced several common troubleshooting techniques for resolving Linux Web App issues caused during the build stage. Most of these problems relate to package compatibility, although the symptoms may vary greatly. By understanding the debugging process demonstrated in these examples, you will be better prepared to diagnose and resolve similar issues in future deployments.Microsoft Defender for Cloud Customer Newsletter
What's new in Defender for Cloud? Defender for Cloud integrates into the Defender portal as part of the broader Microsoft Security ecosystem, now in public preview. This integration, while adding posture management insight, eliminates silos natively to allow security teams to see and act on threats across all cloud, hybrid, and code environments from one place. For more information, see our public documentation. Discover Azure AI Foundry agents in your environment The Defender Cloud Security Posture Management (CSPM) plan secures generative AI applications and now, in public preview, AI agents throughout its entire lifecycle. Discover AI agent workloads and identify details of your organization’s AI Bill of Materials (BOM). Details like vulnerabilities, misconfigurations and potential attack paths help protect your environment. Plus, Defender for Cloud monitors for any suspicious or harmful actions initiated by the agent. Blogs of the month Unlocking Business Value: Microsoft’s Dual Approach to AI for Security and Security for AI Fast-Start Checklist for Microsoft Defender CSPM: From Enablement to Best Practices Announcing Microsoft cloud security benchmark v2 (public preview) Microsoft Defender for Cloud Innovations at Ignite 2025 Defender for AI services: Threat protection and AI red team workshop Defender for Cloud in the field Revisit the Cloud Detection Response experience here.. Visit our YouTube page: here GitHub Community Check out the Microsoft Defender for Cloud Enterprise Onboarding Guide. It has been updated to include the latest network requirements. This guide describes the actions an organization must take to successfully onboard to MDC at scale. Customer journeys Discover how other organizations successfully use Microsoft Defender for Cloud to protect their cloud workloads. This month we are featuring Icertis. Icertis, a global leader in contract intelligence, launched AI applications using Azure OpenAI in Foundry Models that help customers extract clauses, assess risk, and automate contract workflows. Because contracts contain highly sensitive business rules and arrangements, their deployment of Vera, their own generative AI technology that includes Copilot agents and analytics for tailored contract intelligence, introduced challenges like enforcing and maintaining compliance and security challenges like prompt injections, jailbreak attacks and hallucinations. Microsoft Defender for Cloud’s comprehensive AI posture visibility with risk reduction recommendations and threat protection for AI applications with contextual evidence helped preserve their generative AI applications. Icertis can monitor OpenAI deployments, detect malicious prompts and enforce security policies as their first line of defense against AI-related threats. Join our community! Join our experts in the upcoming webinars to learn what we are doing to secure your workloads running in Azure and other clouds. Check out our upcoming webinars this month! DECEMBER 4 (8:00 AM- 9:00 AM PT) Microsoft Defender for Cloud | Unlocking New Capabilities in Defender for Storage DECEMBER 10 (9:00 AM - 10:00 AM PT) Microsoft Defender for Cloud | Expose Less, Protect More with Microsoft Security Exposure Management DECEMBER 11 (8:00 AM - 9:00 AM PT) Microsoft Defender for Cloud | Modernizing Cloud Security with Next‑Generation Microsoft Defender for Cloud We offer several customer connection programs within our private communities. By signing up, you can help us shape our products through activities such as reviewing product roadmaps, participating in co-design, previewing features, and staying up-to-date with announcements. Sign up at aka.ms/JoinCCP. We greatly value your input on the types of content that enhance your understanding of our security products. Your insights are crucial in guiding the development of our future public content. We aim to deliver material that not only educates but also resonates with your daily security challenges. Whether it’s through in-depth live webinars, real-world case studies, comprehensive best practice guides through blogs, or the latest product updates, we want to ensure our content meets your needs. Please submit your feedback on which of these formats do you find most beneficial and are there any specific topics you’re interested in https://aka.ms/PublicContentFeedback. Note: If you want to stay current with Defender for Cloud and receive updates in your inbox, please consider subscribing to our monthly newsletter: https://aka.ms/MDCNewsSubscribe
Best AI Agent for HR in Microsoft Teams?
So our HR team is looking for an AI agent tool that fits into our HR processes, such as performance reviews, 360 feedback, policy management, etc. We've messed around with Copilot and while it has been solid for productivity overall, it doesn't quite handle the HR processes. Any recommendations?
