Announcing AWS For SAP Management MCP Server — Manage Your SAP Applications on AWS Using AI

1 day ago 7

Overview

Checking the health of a production SAP HANA database on AWS today requires 10+ API calls across 6+ AWS services. You check application status in AWS Systems Manager for SAP (SSM for SAP) or AWS Console for SAP Applications, review metrics for CPU, memory and disk in Amazon CloudWatch (CloudWatch), verify HANA backup status in AWS Backup, run configuration checks on the application using SSM for SAP, inspect filesystem usage on SAP hosts, and then correlate all this data manually. Each service has its own AWS Management console, APIs, CLIs, and output formats.
Beyond the manual effort, managing SAP workloads on AWS demands deep knowledge of each service’s APIs, required parameters, and response structures. Building this kind of cross-service workflow yourself means writing and maintaining custom scripts, handling error cases across multiple services, and keeping up with API changes. That expertise is often concentrated on a few team members, creating bottlenecks when they’re unavailable.
The AWS For SAP Management MCP Server simplifies this complexity into a single natural language conversation. Instead of navigating multiple AWS Management consoles, or memorizing different API or CLI input and output structures, you ask your AI assistant (like Kiro—an AI-powered IDE, Kiro-CLI – brings the power of AI-assisted development to your terminal, or others that support MCP server integration), a question related to managing your SAP landscape, and get a structured, SAP-aware answer.
In this post, you will learn how to setup this MCP server, to gain AI-driven insights, execute SAP management automation, and integrate with your existing workflows and AWS environment.

What is an MCP server, and why is it needed for SAP operations

The Model Context Protocol (MCP) is an open standard that lets AI assistants call external tools in a structured way. The MCP server is a specialist toolkit that an AI assistant can use. In this case, a toolkit that understands SAP.
SAP applications on AWS run core business processes like finance, supply chain, and human resources. Today, you manage them through purpose-built experiences delivered through SSM for SAP and AWS Console for SAP Applications. However, without an MCP server, an AI assistant can only work with what it knows, creating a gap between AI capabilities and SAP-specific operations.
The AWS For SAP Management MCP Server brings this capability to AI assistants. It combines what the server knows—SAP component dependencies, validated patterns, and cross-service logic—into one interface you can talk to. By configuring this MCP with your AI assistants, they gain the right SAP context, and access to 20+ SAP-aware tools. These tools can run management queries on your live SAP environment, correlate data across AWS services, return SAP-meaningful results, and execute actions with your approval.

How it works

The AWS For SAP Management MCP Server runs locally on your desktop or existing infrastructure, connecting to your AI assistant via the MCP standard (stdio transport). It uses your existing AWS credentials (~/.aws/config) to make API calls, and no additional infrastructure is required. This MCP server builds on SSM For SAP and uses the registered applications as its foundation. You can register the SAP applications through the AWS Console for SAP Applications, via the SSM for SAP API / CLI or using the AWS API MCP Server by providing the applicable inputs)

Capabilities and safety controls

Published through AWSLabs, this MCP server can be used for different troubleshooting, monitoring, or automation use cases. It understands the SAP application topology — which components make up a system, how they relate to each other, and which AWS resources support them. The MCP server exposes 20+ structured tools across four areas:

Monitoring and reporting (read-only)
- View your full SAP landscape, including all registered applications, component topology, and metadata
- Get health summaries that correlate SAP application status from SSM for SAP, metrics from Amazon CloudWatch, backup status from AWS Backup, filesystem usage, and configuration compliance into a single response
- Generate downloadable Markdown health reports for compliance or team reviews
Configuration compliance (read-only)
- Run configuration checks against the SAP Lens of the AWS Well-Architected Framework (containing best practices to run SAP workloads on AWS)
- Drill into individual rule evaluations showing actual vs. expected values, with remediation guidance
Application lifecycle management (requires confirmation)
- Start and stop SAP applications through natural language
- The server understands component dependencies. When stopping a HANA database, it discovers dependent NetWeaver applications and offers to cascade-stop them in the correct order, along with the underlying EC2 infrastructure, with explicit confirmation before proceeding
- Register new SAP applications with SSM for SAP
Schedules operations (requires confirmation and creates resources)
- Create recurring schedules through Amazon EventBridge Scheduler (EventBridge Scheduler). For example, “Schedule my dev SAP system to run only on weekdays”
- The server handles EventBridge Scheduler configuration, role creation through IAM, and payload construction
- Useful for cost optimization, for example, schedule non-production systems to run only during business hours

On safety: Read-only operations (monitoring, reporting, compliance checks) execute without confirmation. Operations that change state, such as stopping an application or creating a schedule, require explicit confirmation before the server proceeds. Your AI assistant will present the planned action and ask you to approve it.

Example AI prompts

Here are common tasks that you can perform with this MCP server through natural language:

Getting started

Getting started with the AWS For SAP Management MCP Server takes just a few minutes.

Prerequisites

AWS credentials configured via ~/.aws/config with permissions for SSM for SAP, Amazon CloudWatch, Amazon EC2, Amazon EventBridge Scheduler, AWS Backup, SSM, and IAM
SAP applications registered with SSM for SAP are used as a foundation by the MCP sever.
An MCP-compatible AI assistant such as Kiro, Kiro-CLI, and so on.
Python 3.10+ with uvx installed. Use pip install uvx or an equivalent command for your OS, to install uvx and run uvx –version to ensure that it is in your PATH

Installation

Before you begin, ensure you have completed the prerequisites listed above.

Locate your AI assistant’s MCP configuration file (check Kiro documentation or your assistant’s documentation for the file location).
Open the configuration file, and Add the following block to the mcpServers section:

NOTE: If your configuration file doesn’t exist, create it with the complete structure shown above. If it exists but lacks an mcpServers section, add one.”

Replace <YOUR_AWS_PROFILE> with your AWS profile name.
Save the configuration file. You may need to provide the full path to uvx, if it isn’t in your PATH.
Restart your AI assistant to load the new configuration.

Using both MCP servers together

The AWS For SAP Management MCP Server is designed to work alongside the AWS API MCP Server as complementary tools. It handles SAP-specific workflows with domain knowledge, understanding your SAP topology, component dependencies, and operational best practices. The AWS API MCP Server provides general-purpose access to AWS service APIs for operations not covered by the SAP-specific tooling.
Together, they give your AI assistant both SAP-aware intelligence and broad AWS service coverage. We recommend configuring both:

{ "mcpServers":{ "awslabs.aws-for-sap-management-mcp-server":{ "autoApprove":[], "disabled":false, "command":"uvx", "args":[ "awslabs.aws-for-sap-management-mcp-server@latest" ], "env":{ "AWS_PROFILE":"<YOUR_AWS_PROFILE>", "FASTMCP_LOG_LEVEL":"ERROR" }, "transportType":"stdio" }, "awslabs.aws-api-mcp-server":{ "command":"uvx", "args":[ "awslabs.aws-api-mcp-server@latest" ], "env":{ "AWS_PROFILE":"<YOUR_AWS_PROFILE>" }, "transportType":"stdio" } } }

Verify it is working

Open your AI assistant.
Verify that the AWS For SAP Management MCP is connected successfully, and its tools appear in your assistant’s available tools
Test using a simple prompt: “List all my SAP applications” and verify the response matches your expected SAP landscape.
If you see an error, then check the Troubleshooting section below.

Troubleshooting

MCP server not appearing in AI assistant

Verify the configuration file is a valid JSON
Restart your AI assistant after saving configuration changes
Confirm that uvx is installed and added to your PATH (or use the full uvx path in your config).

Authentication errors

Verify your AWS_PROFILE is correctly configured in ~/.aws/config
Confirm the profile has the required IAM permissions

SAP applications not discovered

Verify your SAP applications are registered with SSM for SAP
Check that the AWS Region in your profile matches where your SAP applications are registered

Availability and Pricing

The AWS For SAP Management MCP Server is available today as an open-source project in the AWS Labs MCP repository on GitHub. It is distributed through PyPI, Amazon Elastic Container Registry (Amazon ECR), Docker Hub, and the MCP Toolkit.

The MCP server itself is free. It runs locally on your machine. Standard AWS service usage charges apply for the underlying API calls (SSM for SAP, Amazon CloudWatch, Amazon EventBridge Scheduler, AWS Backup, Amazon EC2), the same charges you would incur using the AWS Management Console or AWS CLI directly.
The server works in AWS Regions, where AWS Systems Manager for SAP is supported.

Conclusion

The AWS For SAP Management MCP Server reduces the operational complexity of managing SAP workloads on AWS. Tasks that previously required navigating multiple consoles and correlating data across services, such as health checks, compliance validation, lifecycle management, and scheduling, become single natural language interactions.

It complements the AWS Console for SAP Applications (visual management) and the SSM for SAP APIs (programmatic access) as a third option: conversational, SAP-aware operations management.

Get started by visiting the AWS Labs MCP repository and configuring the server with your AI assistant. For detailed guidance on AWS Systems Manager for SAP, visit the SSM for SAP documentation and API reference. To learn more about running SAP workloads on AWS, visit AWS for SAP.

Read Entire Article