Overview

The Mg MCP Server is a lightweight Model Context Protocol (MCP) implementation designed to expose OpenShift Must‑Gather data as an AI‑friendly resource. Must‑Gather is a utility that collects diagnostic information from OpenShift clusters, producing a comprehensive tarball of logs, configuration files, and system state. By turning this output into an MCP server, developers can query the data through a standard AI interface, enabling assistants like Claude to retrieve, filter, and interpret cluster diagnostics on demand.

Problem Solved

In large Kubernetes or OpenShift environments, troubleshooting often requires manual extraction of logs and configuration files from multiple nodes. This process is time‑consuming, error‑prone, and difficult to automate. The Mg MCP Server bridges this gap by providing a single, consistent API that AI assistants can call to obtain any piece of Must‑Gather data. It removes the need for custom scripting, streamlines incident response, and reduces the learning curve for operators who may not be fluent in command‑line tooling.

Core Functionality

• Resource Exposure: The server publishes Must‑Gather artifacts (e.g., , ) as MCP resources, allowing clients to request specific files or directories.
• Tool Integration: Built‑in tools parse and format the data, returning human‑readable summaries or structured JSON that AI assistants can consume directly.
• Prompt Templates: Pre‑defined prompts guide the assistant to ask clarifying questions, ensuring that the correct diagnostics are retrieved.
• Sampling & Pagination: For large log files, the server supports streaming and pagination, preventing memory overload while still providing complete context.

Use Cases

• Automated Incident Response: An AI assistant can fetch the latest Must‑Gather tarball, parse relevant logs, and suggest remediation steps without manual intervention.
• Compliance Audits: By exposing configuration files as resources, auditors can query the server to verify that cluster settings meet regulatory standards.
• Developer Onboarding: New engineers can ask the assistant to explain specific configuration sections, receiving instantly generated explanations from the Must‑Gather data.

Integration into AI Workflows

Developers embed the Mg MCP Server in their CI/CD pipelines or observability stacks. AI assistants query it via the standard MCP endpoint, receiving structured data that can be fed into downstream natural‑language generation models. Because the server adheres to MCP specifications, any compliant client—whether a custom UI or a third‑party assistant—can interact with it without bespoke adapters.

Unique Advantages

• Zero Configuration for End Users: The server automatically maps Must‑Gather directories to MCP resources, eliminating manual mapping.
• High Performance: Lightweight Rust implementation ensures low latency even when serving large tarballs.
• Extensibility: Developers can add custom tools or prompts to tailor the assistant’s responses for specific operational contexts.

In summary, the Mg MCP Server transforms static OpenShift diagnostic dumps into a dynamic, AI‑accessible knowledge base, empowering developers and operators to leverage conversational assistants for rapid troubleshooting, compliance checks, and operational insight.