Open Cluster Management MCP Server

The Open Cluster Management (OCM) MCP Server bridges generative AI assistants with the full breadth of a Kubernetes multi‑cluster environment. By speaking the Model Context Protocol, Claude or other AI agents can discover, query, and manipulate resources across a hub cluster and all managed clusters without leaving the chat interface. This eliminates the need for manual sessions or custom scripts, allowing developers to ask high‑level questions—such as “What pods are running on cluster B?” or “Show me the latest alert from the monitoring stack”—and receive precise, context‑aware responses.

At its core, the server exposes a rich set of MCP tools that enable cluster‑aware operations. These include retrieving resources from the current context (the hub) or any managed cluster, establishing connections to a target cluster via a specified , and pulling metrics, logs, or alerts from integrated monitoring stacks. The tool set also supports interactions with OCM‑specific APIs like Managed Clusters, Policies, and Add‑ons, giving agents the ability to not only read but also modify cluster state—deploying a new policy or scaling an application can be scripted through the same conversational interface.

Complementing the tools are prompt templates designed for common OCM tasks. These reusable prompts help standardize agent behavior, ensuring consistent phrasing and context handling when orchestrating multi‑cluster operations. Alongside the templates, a collection of MCP resources links directly to official OCM documentation and related references, enabling the AI to surface authoritative guidance or troubleshooting steps within the conversation.

Real‑world use cases abound: a DevOps engineer can ask the assistant to list all pods that exceed memory thresholds across every cluster, receive aggregated metrics, and automatically trigger a scaling policy; an SRE can request the latest alerts from Prometheus and have them correlated with specific managed clusters, all while maintaining a single source of truth. In CI/CD pipelines, the server can be invoked to validate that new deployments adhere to cluster‑level policies before promotion. Because the MCP server treats the configured as the hub, it seamlessly navigates the OCM hierarchy, abstracting away complex network routing or authentication details.

What sets this MCP server apart is its tight integration with the OCM framework and its focus on interactive observability. The ability to pull real‑time logs, metrics, and alerts into a conversational AI workflow empowers teams to troubleshoot faster, automate compliance checks, and orchestrate cross‑cluster changes with confidence—all through natural language. This streamlined, protocol‑driven approach removes the friction traditionally associated with multi‑cluster management and positions AI assistants as first‑class collaborators in Kubernetes operations.