A man studies a blue cloud infrastructure dashboard displaying secure servers, network connections, and performance charts.
Microsoft has put Azure Workload Center for SAP and the Azure Copilot Run SAP Agent into limited preview. The announcement came on November 17 at Ignite. It is aimed at teams that run SAP on Azure and are tired of piecing together VM metrics, alerts and SAP knowledge by hand.

The preview does not cover every Azure customer. Visual Studio Magazine's Ignite roundup sorts the status of each agentic feature. It lists the deployment agent and Azure Monitor health models as generally available, and Copilot spaces, custom skills and the new SRE Agent capabilities as public preview. Azure Workload Center is in limited preview.

A man studies a blue cloud infrastructure dashboard displaying secure servers, network connections, and performance charts. What Microsoft announced​

Microsoft's Tech Community post says that Azure Workload Center extends agentic operations to Oracle and SAP. The SAP half is a rebrand and expansion of an existing product. The Tech Community post says Azure Center for SAP solutions (ACSS) is now Azure Workload Center for SAP.

ACSS had already made SAP a top-level workload in Azure. It supported guided deployment and registration of existing systems. It represented each system as a Virtual Instance for SAP solutions (VIS), with health and status, quality checks, application-tier start and stop, infrastructure metrics and cost analysis.

The new product shifts the focus to "day-two" operations: reliability, health, best-practice checks and agent-assisted work. Microsoft lists three pillars:

  • Configuration validation against industry best practices, with deviation detection.
  • Faster detection and remediation through workload health that includes application signals, plus guided remediation.
  • Broader coverage of SAP system stacks and architectures.

The Workload Virtual Instance​

The core new idea is the Workload Virtual Instance (WVI). It is a logical collection of the components and Azure resources that make up a workload, such as one SAP system. Each SAP system is enrolled as one WVI.

Child components represent SAP central services, application servers, the database and web dispatchers. Each component maps to its Azure compute, network and storage resources. A WVI can also hold metadata such as the SAP system identifier (SID), product, environment and high-availability architecture.

That metadata is what the agent works from. An AI assistant that knows a VM is the HANA database node in a production, HA-clustered system can reason differently from one that only sees "a VM with high CPU." The Azure Updates entry, as syndicated by Cloud Changelog, says SIDs and production designations help teams and AI agents understand each environment in context. It adds that Azure Monitor health models provide health insights across the application, database and infrastructure layers.

Microsoft also says Azure Workload Center for Oracle is in limited preview. Both workload centers are built on the WVI.

Onboarding and supported landscape​

Microsoft says registration now extends to Java stack systems, disaster recovery systems, Web Dispatchers and SAP HANA scale-out architectures. Onboarding works in two ways:

  • ABAP-based systems: automated discovery through the SAP Host Control API, or Azure Resource Graph, which finds supported systems and creates a WVI for each.
  • Java-based systems: Azure Resource Graph, or manual onboarding, where you define the Azure resources for each component.

Both methods are available through the Azure portal and APIs. Microsoft's post gives no compatibility matrix, minimum versions, permission requirements or step-by-step procedure. Check those details before planning a rollout.

Configuration validation​

Microsoft's argument is that SAP environments rest on hundreds of configuration decisions across Azure infrastructure, operating systems, clustering and SAP components. Settings that were correct at deployment drift over time, and new resources don't always follow the same standards. The product checks systems against Microsoft and SAP best practices and flags deviations.

The preview adds more checks. You can view all of them and choose which ones run through configuration templates. Results, affected components and recommended fixes appear in the portal, or you can ask the agent to summarize them. Notifications are consolidated by system, which should cut repetitive alerts.

The workflow Microsoft describes runs like this:

  1. Spot a deviation in the portal.
  2. Ask the Run SAP agent what changed and what the impact is, and have it retrieve current settings.
  3. Review the workload-aware remediation plan and generated scripts.
  4. Decide when and where to run the changes.
  5. Confirm the fix on the next validation run.

Microsoft is explicit that the agent analyzes, describes and recommends, and can generate remediation scripts. The customer decides when and where any change runs. The announcement does not describe autonomous remediation.

Health and troubleshooting​

Workload Health Observability is built on Azure Monitor health models. It combines infrastructure metrics with optional application signals and rates the whole workload as healthy, degraded or unhealthy. A newly onboarded WVI comes with predefined infrastructure metrics, thresholds and health logic, and you can customize them.

You can add application signals such as SAP HANA availability and set alerts at the workload or component level. A dashboard timeline, dependency graph and resource-level views show when health changed and which components were affected. An AI-generated summary points to issues that need attention and where to start investigating.

For outages and slowdowns, Impact Analysis starts with Azure resources, initially virtual machines. It brings in diagnostic signals and correlates Azure platform signals. You can ask in natural language for:

  • The probable root cause.
  • The affected VMs and other resources.
  • The scope, including other VMs hit by the same cause.

What isn't here yet​

Microsoft says it plans to extend investigation to SAP performance and availability issues. That would correlate SAP application signals with Azure platform and infrastructure evidence. This is roadmap, not a shipped capability. Today's Impact Analysis is about Azure resources, starting with VMs.

Microsoft also says the product is meant to complement existing monitoring platforms and incident-management processes, not replace them.

Availability, cost and access​

Microsoft's announcement lists these regions: Australia East, Central US, East US, East US 2, Germany West Central, Italy North, Japan East, Korea Central, North Europe, South Central US and West Europe. The Cloud Changelog copy of the Azure Updates text omits West Europe, so confirm your region when you sign up.

Microsoft says there is no additional licensing fee during the limited preview. Customers pay only for the underlying infrastructure deployed as part of the product or the services they enable. Access is through a sign-up form, and Microsoft hasn't said how it picks participants.

Analysis: who should care, and what to watch​

The pitch is real operational pain. Correlating Azure signals, SAP signals and configuration state across layers is slow and error-prone. A shared workload model is a sensible foundation.

The claims are still Microsoft's. The announcement contains customer and partner endorsements from TCS and Protera, but no measured time savings or outcomes. Microsoft's own Azure Updates page also separates "in preview" from "in development." It defines the latter as a limited offering for select customers, for non-production use and testing. Treat this as a trial, not something to put in a production runbook yet.

Sources also describe the preview differently. A WindowsForum Ignite report notes that Microsoft's Book of News calls Workload Center a private preview, while Microsoft's Azure roundup says limited preview. The Tech Community post uses "limited preview."

If you run SAP on Azure, a sensible approach is:

  • Apply to the preview with a non-production landscape first.
  • Test how well your Java, DR or HANA scale-out systems onboard.
  • Compare the built-in configuration checks against what you already run, such as Azure Advisor-based quality checks in ACSS.
  • Review agent-generated scripts as you would any change from a junior admin: carefully, in staging, before production.
  • Check how the preview fits your Entra permissions and change-control policy. Microsoft's post doesn't cover those details.

The broader trend is clear. Microsoft is building agents that sit on top of structured workload context, not generic chat over logs. The WVI is that context layer for SAP and Oracle. Whether it earns trust will depend on how accurate the checks are and how well the agent's remediation plans hold up in complicated SAP landscapes.

 

References

  1. (In preview) Limited Preview: Azure Workload Center for SAP and Azure Copilot Run SAP Agent Azure Updates 2026-11-17T22:48:10Z
  2. Announcing Limited Preview of Microsoft Azure Workload Center for SAP | Microsoft Community Hub techcommunity.microsoft.com
  3. Limited Preview: Azure Workload Center for SAP and Azure Copilot Run SAP Agent — Cloud Changelog cloudchangelog.com