Red Hat Advanced Cluster Management for Kubernetes (RHACM) 2.17 adds a Technology Preview Model Context Protocol (MCP) server that lets AI assistants query the product’s fleet-wide resource search. The release also brings cluster context to OpenShift Lightspeed and a handful of GitOps and observability previews.
Fleet search as an MCP tool
The new server reads the same search database RHACM uses to index Kubernetes resources across managed clusters. Its find_resources tool filters by kind, name, namespace, cluster, status, labels, cluster selectors, text and age, and can return lists, counts, summaries or health views. Red Hat documents connecting it to OpenShift Lightspeed (added through the cluster’s OLSConfig) and to Claude Code, which authenticates with an OpenShift token.
It’s installed with a Helm chart on the RHACM hub cluster, needs cluster-admin access, and is labeled Technology Preview, so it’s meant for evaluation rather than supported production use.
Other changes in 2.17
- Lightspeed context: from a managed cluster’s details page, the console can attach YAML from the
ManagedClusterandManagedClusterInforesources to a Lightspeed conversation. - Argo CD agent mode (preview): uses destination-based mapping for
ApplicationSetdeployment with less up-front configuration. - Perses dashboards (preview): hub and managed-cluster metrics can render through the Cluster Observability Operator console plug-in.
- Governance: policies can detect available minor operator updates on other channels.
- Virtualization: admins can distribute custom roles to managed clusters and assign them with
MulticlusterRoleAssignment.
Why it matters
For teams running many OpenShift clusters, the MCP server means an assistant can answer questions grounded in the actual state of the fleet instead of only generic documentation. Because it’s a preview, it’s best treated as something to test in inventory and troubleshooting workflows for now.
Source: upstreambeat, summarizing Red Hat’s RHACM 2.17 release notes.