rosa
MCP server from redhat-community-ai-tools/rosa-mcp-server
claude mcp add --transport stdio redhat-community-ai-tools-rosa-mcp-server ./rosa-mcp-server --transport=stdio \ --env OCM_OFFLINE_TOKEN="Set your OCM offline token here for stdio transport"
How to use
ROSA MCP Server exposes a set of 6 core tools to manage ROSA HCP clusters and prerequisites via a Model Context Protocol (MCP) interface. It supports both stdio and Server-Sent Events (SSE) transports, enabling local development and remote streaming integrations. With OCM API integration and multi-region support, you can list clusters, inspect specific clusters, create ROSA HCP clusters, retrieve prerequisites guidance, and set up an HTpasswd identity provider for cluster access. To start with stdio transport (local development), set the OCM_OFFLINE_TOKEN environment variable and run the server in stdio mode. For remote usage, run in SSE mode and pass authentication via the X-OCM-OFFLINE-TOKEN header in requests. The tools available include: whoami for account details, get_clusters to filter clusters by state, get_cluster for details of a given cluster, create_rosa_hcp_cluster to provision clusters with required AWS configuration, get_rosa_hcp_prerequisites_guide to obtain the complete prerequisites workflow, and setup_htpasswd_identity_provider to enable username/password login for a cluster.
How to install
Prerequisites:
- Go 1.21 or later
- OCM offline token from console.redhat.com/openshift/token
Build from Source:
git clone https://github.com/tiwillia/rosa-mcp-go
cd rosa-mcp-go
make build
Container Installation:
# Build container image
make container-build
# Run with SSE transport
make container-run
OpenShift Deployment:
# Deploy to OpenShift
make deploy
# Configure template parameters as needed:
# - IMAGE: Container image (default: quay.io/redhat-ai-tools/rosa-mcp-server)
# - MCP_HOST: Hostname for the route
# - CERT_MANAGER_ISSUER_NAME: TLS certificate issuer
Configuration (CLI and TOML):
rosa-mcp-server [flags]
# Flags include:
# --config string path to configuration file
# --host string host for SSE transport (default "0.0.0.0")
# --ocm-base-url string OCM API base URL (default "https://api.openshift.com")
# --ocm-client-id string OCM client ID (default "cloud-services")
# --port int port for SSE transport (default 8080)
# --sse-base-url string SSE base URL for public endpoints
# --transport string transport mode (stdio/sse) (default "stdio")
# TOML example
ocm_base_url = "https://api.openshift.com"
transport = "stdio"
port = 8080
sse_base_url = "https://example.com:8080"
Additional notes
Environment and configuration tip: For stdio transport, you must provide OCM_OFFLINE_TOKEN in the environment. For SSE transport, authentication is header-based via X-OCM-OFFLINE-TOKEN; no environment variable is required. The default ROSA region is us-east-1, but you can configure regional deployment via the config or flags. If you encounter issues deploying to OpenShift, ensure your AWS IAM roles (installer, support, worker) and ROSA creator ARN are correctly configured, and that you provide valid subnet IDs and availability zones. The HBHTPASSWD identity provider setup expects a cluster_id and a list of user:password pairs; ensure the target cluster is accessible and that you have permission to modify identity providers for that cluster.
Related MCP Servers
trpc-agent-go
trpc-agent-go is a powerful Go framework for building intelligent agent systems using large language models (LLMs) and tools.
station
Station is our open-source runtime that lets teams deploy agents on their own infrastructure with full control.
tiger-cli
Tiger CLI is the command-line interface for Tiger Cloud. It includes an MCP server for helping coding agents write production-level Postgres code.
gopls
MCP server for golang projects development: Expand AI Code Agent ability boundary to have a semantic understanding and determinisic information for golang projects.
kubernetes
A Model Context Protocol (MCP) server for the Kubernetes API.
gcp-cost
💰 An MCP server that enables AI assistants to estimate Google Cloud costs, powered by Cloud Billing Catalog API and built with Genkit for Go