Frequently Asked Questions about the Grafana MCP Server

Q: What is the Model Context Protocol (MCP)? A: MCP (Model Context Protocol) is an open standard that standardizes how applications provide context to LLMs (Large Language Models), enabling them to access and interact with external data and tools.

Q: What is the purpose of the Grafana MCP Server? A: The Grafana MCP Server acts as a bridge, allowing AI models and agents to access and interact with Grafana data and functionalities. It translates AI Agent requests into Grafana API calls, allowing agents to understand metrics, dashboards, and incidents.

Q: What Grafana functionalities can AI Agents access through the MCP Server? A: AI Agents can access functionalities like searching for dashboards, retrieving dashboard details, querying datasources (Prometheus, Loki), exploring Prometheus and Loki metadata, managing incidents, and interacting with Grafana OnCall.

Q: What are some use cases for the Grafana MCP Server? A: Use cases include automated anomaly detection, intelligent troubleshooting, proactive performance optimization, automated capacity planning, and AI-powered chatbots for observability.

Q: How do I configure the Grafana MCP Server? A: You need to create a service account in Grafana, download the server, configure it with your Grafana URL and API key, and then configure your AI Agent to connect to the server using the Model Context Protocol.

Q: What is UBOS, and how does it enhance the Grafana MCP Server? A: UBOS is a full-stack AI Agent Development Platform. UBOS provides a centralized asset marketplace, simplified agent orchestration, secure data access, and comprehensive monitoring and logging. It streamlines the development and management of AI Agents that utilize the Grafana MCP Server.

Q: Can I customize the Grafana MCP Server? A: Yes, you can customize the server by adding custom tools, modifying existing tools, and integrating with custom AI Agents. UBOS provides tools and resources for customization.

Q: What are the benefits of using the Grafana MCP Server? A: Benefits include enabling AI Agents to understand real-time data, automate tasks, improve troubleshooting, optimize performance, and drive better business outcomes.

Q: What kind of access permissions are required for the Grafana service account? A: The Grafana service account needs enough permissions to use the tools you want to expose to the MCP client. Refer to the Grafana documentation for details.

Q: How can I troubleshoot connection issues between Claude Desktop and the Grafana MCP Server? A: If you see Error: spawn mcp-grafana ENOENT in Claude Desktop, specify the full path to mcp-grafana in your configuration.

Q: How can I enable debug mode to troubleshoot issues between the MCP server and the Grafana API? A: Enable debug mode by adding the -debug flag to the mcp-grafana command. This will provide detailed logging of HTTP requests and responses.