Connects Claude to Prometheus monitoring through 11 tools spanning instant queries, range queries, and cluster diagnostics. Built in Go with native Kubernetes integration via client-go, so it can automatically proxy through the K8s API to reach Prometheus without port-forwarding. You get the standard PromQL operations plus higher-level tools like getClusterHealthOverview, diagnoseNode, and getTopResourceConsumers that bundle common observability patterns. Works with direct Prometheus URLs or auto-detects OpenShift monitoring setups. Reach for this when you want Claude to query metrics, investigate pod issues, or compare performance across time windows without leaving the conversation.
claude mcp add --transport stdio jeanlopezxyz-mcp-prometheus -- npx -y mcp-prometheus