The Trento MCP Server is a Model Context Protocol (MCP) implementation that bridges the gap between AI assistants and the Trento Project. Instead of navigating through web interfaces, you can now manage and monitor your SAP systems by simply conversing with your AI assistant in natural language.

Whether you're checking cluster health, reviewing system configurations, or analyzing SAP landscapes, the Trento MCP Server enables you to work more efficiently by bringing Trento's powerful capabilities directly into your AI-powered workflow. It connects to SUSE Linux Enterprise Server (SLES), which provides native agentic AI support through MCPHost integration, and it is also supported by SUSE AI for private, on-premises AI deployments.

Watch this video demonstrating how the Trento MCP Server transforms complex SAP monitoring tasks into simple natural language conversations.

Usage

The officially supported ways to run Trento are described at the SUSE documentation website. It covers the RPM and the Kubernetes-based deployments.

If you want to run Trento using containers, please refer to the Helm chart available at the Trento Helm Charts repository. It is the supported way to run Trento using containers in a Kubernetes environment, as it takes care of all the necessary dependencies and configurations for you.

You can install the Trento Helm Chart using the following command:

helm upgrade \
   --install trento-server oci://registry.suse.com/trento/trento-server \
   --create-namespace \
   --namespace trento \
   --set global.trentoWeb.origin=YOUR_TRENTO_SERVER_HOSTNAME \
   --set trento-web.adminUser.password=YOUR_ADMIN_PASSWORD \
   --set trento-mcp-server.enabled=true

Refer to the official SUSE documentation for detailed instructions on configuring the Trento MCP Server and connecting it to your AI assistant.

Once configured, you can interact with Trento through your AI assistant using natural language:

Ask questions conversationally: "Show me all SAP systems in my landscape" "What's the health status of cluster cluster-1?" "List all hosts running HANA databases" "Get details about the checks execution history" "Are there any critical alerts I need to address?"
Get instant insights: "Summarize the overall health of my SAP environment" "Which systems need attention today?" "Show me the latest check results for production systems"

The AI assistant will use the Trento MCP Server to execute these requests and present the results in a conversational format.

Using the container image directly

Running the container image directly is not the recommended way to run Trento and it is not supported by SUSE. It requires manual configuration of all dependencies and environment variables, including PostgreSQL and RabbitMQ.

If you still want to run Trento from the container image without Kubernetes, please refer to the project documentation.

As a quick example, you can run the container image using the following command:

docker run -d \
  --name mcp-server-trento \
  -p 5000:5000 \
  -e TRENTO_MCP_TRENTO_URL=https://demo.trento-project.io/ \
  registry.suse.com/trento/mcp-server-trento

Licensing

SPDX-License-Identifier: Apache-2.0

This project is licensed under the Apache License 2.0. See the LICENSE file for details.