Overview
An MCP (Model Context Protocol) server application for Kubernetes operations that provides a standardized API to interact with multiple Kubernetes clusters simultaneously using multiple kubeconfig files. The server supports configuration via MCPO's config.json, where you define an mcpServers.kubernetes entry with the command, arguments, and a KUBECONFIG_DIR environment variable. All kubeconfig files placed in that directory represent individual clusters, enabling cross-cluster management from a single interface. Context selection is achieved by specifying the cluster context in commands, allowing seamless switching between clusters. The MCP server exposes a comprehensive set of tools across core, API discovery, metrics, rollout management, scaling, and resource handling to list contexts or namespaces, retrieve resources, monitor usage, manage deployments, apply changes, patch resources, and diagnose issues. It includes capabilities for node and pod operations, resource creation and labeling, and workload exposure. The README provides usage examples for multi-cluster operations and diagnoses, along with installation steps via Smithery, prerequisites (Python 3.8+, pip, uv), virtual environment setup, dependency installation, environment configuration, and launching app.py.
Features
Core Cluster Discovery and Inspection
List available contexts, namespaces, and nodes; retrieve resources; describe resources; and discover APIs and CRDs across clusters.
Cross-Cluster Resource Visibility
List and inspect Kubernetes resources across multiple clusters and contexts from a single interface.
API Discovery
Discover available APIs and Custom Resource Definitions (CRDs) within clusters.
Metrics and Telemetry
Display resource usage for nodes and pods; diagnose issues via status, events, and logs.
Rollout Management
Get rollout status and history; undo, restart, pause, and resume rollouts.
Scaling and Autoscaling
Scale resources and configure Horizontal Pod Autoscalers (HPA).
Resource Creation and Modification
Create resources from YAML/JSON, apply changes, patch resources, and update labels/annotations.
Node and Pod Operations
Cordon/uncordon/drain/taint nodes; execute commands in pods.
Who Is This For?
- DevOps:Manage and compare multiple Kubernetes clusters from a single interface using multiple kubeconfigs.
- SREs:Coordinate cross-cluster reliability and issue diagnosis across dev/stage/prod environments.
- Platform teams:Enable centralized multi-cluster Kubernetes operations for platform teams.
