One place to see it, one agent to act on it
Resiliency is easy to promise and hard to prove. You need to know where your applications are exposed, what to fix first, and whether recovery will work when it counts. This year, we set out to make all three simpler.
At Microsoft Build, we introduced Azure Infrastructure Resiliency Manager in public preview, a goal-driven, application-centric way to assess and improve resiliency. In September, we added the resiliency agent in Azure Copilot, also in public preview. It brought assessment, prioritized recommendations, and deployment-ready code into a single conversation, so you could move from knowing what needs fixing to fixing it.
Then you told us what was still missing. Resiliency signals were scattered, with zone resiliency, disaster recovery, and backup each in a separate view. And the agent stopped short of the finish line. It helped you Start Resilient and Get Resilient, but Stay Resilient was still to come.
At Microsoft Ignite 2026, we’re closing both gaps. Azure Infrastructure Resiliency Manager is now generally available, and its zonal resiliency experience has grown from 20 to 41 Azure services. A new converged resiliency view in the Azure portal brings infrastructure and data resiliency together, with the agent one click away from every insight. The agent now covers the full infrastructure resiliency lifecycle, from an architecture diagram to a validated recovery, and it brings data resiliency with Azure Backup into the same conversation.
What's new at a glance
- A converged resiliency view of infrastructure and data resiliency, with contextual top actions that help you get started with Resiliency Agent.
- General availability of Azure Infrastructure Resiliency Manager, with zonal resiliency coverage expanded from 20 to 41 Azure services.
- Start resilient from an architecture diagram, all the way to validated Bicep, ARM or Terraform templates.
- Get resilient with automatic discovery of your resources and their dependencies, validation of your existing IaC templates, cost- and downtime-aware assessments, and a shareable posture report.
- Stay resilient with SLIs and error budgets, agent-built recovery plans, and drills that prove recovery, with native faults now covering 20 Azure services, up from 9.
- Data resiliency with Azure Backup enablement, backup as code, and protection insights in the same conversation.
Converged view across infrastructure and data resiliency
The converged resiliency view is where your resiliency journey starts. The "getting started" page unifies infrastructure, and data resiliency in one surface in the Azure portal, so you can see where you're exposed before you decide what to fix.
- For your applications, go to ‘Service group resiliency’ and see application criticality goal coverage, resilience state, recovery plan and drill status, and top recommendations, with each service group's zonal state, last disaster recovery drill, and backup coverage side by side.
- For your resources, see what percentage is zonally resilient, what percentage is protected by Azure Backup, and which resources need attention first by going to the ‘Resource resiliency’ tab.
- From there, go deeper into infrastructure resiliency (goals, recommendations, recovery plans, and drills) or Backup & Recovery (protection coverage, jobs, alerts, security posture, and threat detection).
Every insight comes with a next step. Top actions such as "you have "service groups” with unassigned "goals" and "you have resources with no backup protection" open the resiliency agent with a relevant prompt already filled in. The view shows you where you're exposed; the agent helps you close the gap.
Infrastructure resiliency: Start, Get, and Stay Resilient
Application resiliency is a continuous journey, not a one-time task, and it runs in three phases. When the agent launched, it covered the first two. With this update, it supports all three: build resiliency in from the first design, close the gaps in what you already run, and keep proving that your applications can recover.
Start resilient: from whiteboard to zone-resilient code
Most new applications begin life as a diagram, and now you can start from that diagram directly. Upload it, and the agent identifies the resources in your design, confirms your preferences, and recommends how to make each one zone resilient, along with the prerequisites, downtime, and cost impact of each change.
When you're ready, ask for Bicep, ARM or Terraform templates. The agent generates validated, deployment-ready files with resiliency built into every resource definition, ready to open in VS Code or take to a GitHub pull request. You don’t need to limit yourself to the portal for this. You can plug in the resiliency agent in your development workflows, as these tools are available from Github Copilot as well.: https://learn.microsoft.com/en-us/azure/copilot/resiliency-agent#use-azure-resiliency-agent-in-github-copilot-limited-preview. It can also help validate your deployment templates. For current service coverage, see: Infrastructure Resiliency Manager support matrix for goals and recommendations | Microsoft Learn
Try prompts like:
- "I have an architecture diagram that I want to use to deploy my application."
- "Generate Bicep templates for the resilient configuration."
The payoff: the diagram your architects already drew becomes the blueprint for resilient infrastructure.
Get resilient: from discovery to decision
For most teams, the bigger exposure is in what's already running. Select "discover and create a service group for your application" in the converged view, and the agent automatically discovers your resources and their dependencies, then groups them into applications for you. That makes the path to a service group much faster than mapping resources and dependencies by hand. With the application modelled, set its criticality tier and resiliency goals, then ask for an assessment. For every resource that isn't zone resilient, the agent explains how to fix it, the downtime involved, and the cost implications, separating simple configuration changes from those that need a migration or redeployment. Download the posture report and share it with key stakeholders across your organization. The agent recommends; your team decides.
Already managing infrastructure as code? Upload an existing template, and the agent assesses it for resiliency gaps and suggests code snippets to close them.
Try prompts like:
- "Help me discover the applications in my subscription."
- "Validate my Terraform template against Azure resiliency best practices."
- "Give me a resiliency assessment for my service group Contoso Payments, including how to fix each gap, the downtime, and the cost."
Stay resilient: measure, plan, and prove recovery
Getting resilient is a milestone, not the finish line. Applications change, health drifts, and a recovery plan that has never been tested is still an assumption. Staying resilient means watching how your application behaves, acting before drift becomes downtime, and proving you can recover.
Observability and measure what matters. Attach a service level indicator (SLI) to a service group's resiliency goal and track its error budget alongside your zonal posture, so drift shows up while there's still time to act.
Know what's at risk. When the error budget runs low, ask the agent what to do. It reassesses the application, rates the risk in each area, and recommends the action that closes it.
Plan the recovery. The agent generates a recovery orchestration plan that recovers your resources in the order your application needs, with validation checkpoints between stages. You confirm the plan before it's created.
Prove it with a drill. The agent configures a drill with Azure-recommended faults for eligible resources; you confirm the setup and run the drill from the Azure portal. Native faults now cover 20 Azure services, up from 9, so more of your application can take part in a drill without custom fault logic. Afterward, the assessment reflects proven recovery behavior, keeping you prepared for any impact to your applications.
For supported resources and faults, see the recovery plans and drills documentation: Support matrix for Azure Recovery Orchestration Plan | Microsoft Learn, Availability zone down drills support matrix for Infrastructure Resiliency Manager | Microsoft Learn
Try prompts like:
- "My error budget is down to 30%. What's at risk, and what should I do?"
- "Create a recovery orchestration plan for my service group."
- "Create a drill for my service group."
The payoff: resiliency stops being a snapshot from your last assessment and becomes something you measure continuously and prove on your own schedule, before an outage does it for you.
With the foundation laid for zonal resiliency, we are now extending service groups to also carry regional resiliency goals alongside zonal ones through the new regional resiliency experience in Azure Infrastructure Resiliency Manager, now in private preview. Read more here: Beyond availability zones - Building resilience across regions with Azure | Microsoft Community Hub
Data resiliency: Azure Backup comes to the agent
Infrastructure resiliency keeps your application running through a failure. Data resiliency makes sure you have a clean copy to recover from, whether you face an outage, an accidental deletion, or a ransomware attack. With Azure Backup now in the resiliency agent, the backup gaps you see in the converged view are one conversation away from being closed.
- Protect VMs end to end. The agent picks or creates the right vault and backup policy, shows you the full plan, and verifies protection once you approve.
- Backup as code. Generate Azure CLI, PowerShell, Bicep, or ARM to enable VM backup.
- Answers about your protection estate. Ask about coverage, jobs, alerts, and security posture across Azure Backup and Azure Site Recovery, and troubleshoot failed jobs.
- Manage and harden. Update backup policies and strengthen vault security, from soft delete and immutability to multi-user authorization.
Every change is confirmed before it's made, irreversible actions need a fresh confirmation, and restores are guided in the Azure portal. For supported workloads and scenarios, see the data resiliency documentation [link to be added].
Try prompts like:
- "Show me the resources in my resource group that aren't protected, and help me protect them."
- "Generate a Bicep template to enable backup for my VM."
- "Improve the security level of my vault from Good to Excellent."
What to expect in the resiliency agent preview
- Templates generated from an architecture diagram are available in Bicep, Terraform and ARM (JSON) formats.
- Assessments, remediation, recovery plans, and drills in the agent focus on zonal resiliency. Some resource types are reported for visibility but aren't yet supported for automated remediation.
- Regional resiliency goals are set and tracked in Azure Infrastructure Resiliency Manager (private preview) and aren't yet part of the agent's assessment.
- The agent creates and configures recovery plans and drills; you run drills from the Azure portal.
- Data resiliency in the agent starts with Azure VM backup. Other workloads, and restores, are guided through the Azure portal.
Start here: open the resiliency page in the Azure portal pick a top action, and let the agent take it from there.
Where we are going
We're building toward end-to-end, agent-led resiliency across infrastructure, data, and cyber. Next up: regional resiliency in the agent and broader service and workload coverage. We're shaping this with you during preview, so let us know what's working and what you'd like next at azureresiliency@microsoft.com.
See it at Microsoft Ignite
LAB721, "Hands-on with the resiliency in Azure: Assess, recommend, recover,"
BRK452 AI-powered resiliency for modern workloads with Mark Russinovich
BRK445 Architecting for end-to-end resiliency on Azure
Resources
Learn more about Azure Infrastructure Resiliency Manager
Learn more about resiliency agent