Rafay Systems Introduces Managed MCP Server for Governed AI Infrastructure Operations

Rafay Systems has introduced a managed Model Control Protocol (MCP) Server to support governed AI assistance for infrastructure operations, announced in a press release. The new server provides a controlled interface that allows platform, DevOps, and site reliability engineering teams to connect AI assistants and agentic applications to Rafay's operational context without exporting data or custom integrations.

The first supported workflows address fleet intelligence and cost attribution. Engineers can use an MCP compatible assistant to review cluster configurations, health, and resource usage, identify underutilized environments, and analyze cost drivers across Kubernetes projects managed through Rafay. This enables teams to consolidate or optimize infrastructure based on actionable insights.

The second workflow focuses on incident diagnosis and operational triage. Through the MCP Server, users can query degraded clusters or namespaces and receive structured diagnoses for common issues such as failed deployments, unschedulable pods, and configuration gaps. The server provides recommendations while teams maintain full control of operational changes.

The Rafay MCP Server is built on the open source Model Control Protocol, originally introduced by Anthropic under the Agentic AI Foundation. It connects AI clients to the Rafay Platform using existing access controls and enforces authorization through Rafay's role-based access control model. Access is currently read only and can be enabled via a feature flag through Rafay Support.