Openshift-mcp-server

List all clusters managed by Red Hat Advanced Cluster Management (ACM)

Get detailed information about a specific ACM managed cluster

List governance policies across managed clusters (ACM)

Search for resources across all ACM managed clusters using the search API

List available Ansible Automation Platform job templates for OpenShift remediation

Launch an Ansible job template to remediate an OpenShift issue (use after ITSM approval)

Check the status of a running Ansible job

List available Ansible inventories

List Ansible workflow templates for multi-step remediation

Launch an Ansible workflow template for complex multi-step remediation

Add or list namespaces being watched for application code/config changes at the pod level

Scan watched namespaces for application changes (image updates, config changes, env var modifications, replica changes)

View the history of detected application changes across watched namespaces

Rollback a workload to its previous baseline state (reverts image, env, replicas to the snapshot before the change)

Check ArgoCD/OpenShift GitOps applications for drift between live state and Git-desired state

Run kube-burner cluster density/stress tests. Creates a Job that runs kube-burner with the specified workload profile.

Run FIO storage performance benchmarks. Tests IOPS, throughput, and latency for a given storage class.

Test network throughput, latency, or packet loss between pods using iperf3.

Run CPU and/or memory stress tests on worker nodes using stress-ng.

Run database performance benchmarks using pgbench (PostgreSQL) or sysbench (MySQL).

Get OpenShift cluster version, status, and configuration details

Get recent cluster events, optionally filtered by namespace or type (Warning/Normal)

Get cluster-wide resource allocation and usage summary (CPU, memory) across all nodes

Run CIS Kubernetes Benchmark and OpenShift hardening checks against the cluster. Returns pass/fail with remediation commands for each control.

Get a comprehensive cluster summary suitable for the MCP AI Assistant dashboard — nodes, pods, deployments, health, alerts

Parse a document (Word or text) describing an application, extract deployment intent, generate Kubernetes/OpenShift manifests, and deploy

Analyze a pod for common issues (CrashLoopBackOff, ImagePullBackOff, OOMKilled, pending scheduling) and suggest fixes

Perform a health check on all workloads in a namespace — find failing pods, unhealthy deployments, quota issues

Perform a comprehensive cluster health check — nodes, operators, etcd, critical pods, certificate expiry

Compare two ACM-managed clusters and detect configuration drift: image version mismatches, resource differences, policy compliance gaps, and security posture differences

Execute an emergency fix: creates an emergency change request in ServiceNow, performs the fix immediately, and sends notification. Use only for critical production issues.

Execute a fix after ITSM change request has been approved. Checks approval status before executing.

Get any Kubernetes resource by apiVersion, kind, and name

List any Kubernetes resources by apiVersion and kind, with optional selectors

Create any Kubernetes resource from a JSON manifest

Delete any Kubernetes resource by apiVersion, kind, and name

Get or update the scale of any scalable Kubernetes resource (Deployment, StatefulSet, ReplicaSet, etc.)

Apply (server-side patch) a Kubernetes resource from a JSON manifest

List all ArgoCD / OpenShift GitOps applications with sync and health status

Get detailed status of an ArgoCD application including sync, health, source, managed resources, conditions, and operation state

Trigger a sync operation for an ArgoCD application to reconcile it with the desired state in Git

Show resources that are out of sync in an ArgoCD application — what differs between the live state and Git

List ArgoCD AppProjects with their allowed source repos, destinations, and roles

Rollback an ArgoCD application to a specific Git revision (SHA or tag) by updating targetRevision and triggering a sync

GPU hardware on this cluster, read from the Kubernetes API: per-node model, count, allocated vs free, memory per GPU, driver version and MIG capability, plus which pods are holding GPUs. Works without Prometheus.

Whether this cluster has GPU hardware and whether the NVIDIA GPU Operator is installed and working. Answers from the API server, so it is valid even when monitoring is broken.

Full GPU fleet view: inventory from the Kubernetes API, enriched with live DCGM utilisation, memory, temperature, power and error counters when the exporter is being scraped.

List installed Helm releases by querying Helm's secret-based storage backend (secrets with label owner=helm)

Install a Helm chart into a namespace. Since Helm chart rendering requires the Helm CLI, this tool records the intent and provides the equivalent CLI command or suggests an Ansible playbook.

Uninstall a Helm release by deleting its release secrets from the namespace

Get the status of a Helm release including its current revision, status, and metadata

Get the revision history of a Helm release showing all past and current versions

Scan deployed container images for security vulnerabilities (CVEs, supply chain risks, image hygiene, compliance). Uses Quay CSO if available, otherwise static analysis. Reports CVSS scores, compliance badges, and image age.

Get detailed vulnerability report for a specific image or view scan history

Check image supply chain compliance — signing, SBOM, digest pinning, registry trust, CIS benchmarks

Predict the impact of a cluster change before executing it: resource pressure, quota usage, PDB violations, and dependency effects

List KubeVirt virtual machines in a namespace (or all namespaces)

Get detailed information about a specific KubeVirt virtual machine

Start a KubeVirt virtual machine by creating a VirtualMachineInstance

Stop a running KubeVirt virtual machine

Restart a running KubeVirt virtual machine

List running KubeVirt VirtualMachineInstances (VMIs) in a namespace or cluster-wide

Provision an OpenShift Virtualization (KubeVirt) virtual machine with a PERSISTENT root disk and cloud-init access. Supports dry-run. Prefer a golden image (sourceDataSource) plus an instanceType/preference over raw cpu/memory.

Report on VMs this platform provisioned: fleet inventory with provenance, VMs past their decommission date, and VMs whose real usage no longer matches the size chosen for them. Read-only — every recommendation carries a change request for a human to approve.

How to connect to a virtual machine: the cloud-init user, reported IP addresses, guest agent state, and the exact virtctl/ssh commands to use.

List what can be provisioned on this cluster: golden image DataSources, VirtualMachineClusterInstancetypes (sizes) and VirtualMachineClusterPreferences (guest OS tuning).

List resource consumption (CPU and memory) for Kubernetes Pods via metrics-server

List resource consumption (CPU and memory) for Kubernetes Nodes via metrics-server

Collect a diagnostic snapshot of the cluster (lightweight must-gather)

Collect diagnostic info for a specific namespace

Check etcd cluster health and performance

Check for expiring certificates in the cluster

Check resource quotas and limit ranges across namespaces

List all namespaces/projects in the OpenShift cluster with status and labels

Get detailed information about a namespace including resource quotas, limit ranges, and workloads summary

Check service endpoints and endpoint slices

List OpenShift routes

List Kubernetes ingresses

List network policies in a namespace

Get CoreDNS / DNS operator configuration

Check if a service is healthy (service + endpoints + pods)

Get detailed info about a specific OpenShift route

List all OpenShift cluster nodes with status, roles, and resource capacity

Get detailed information about a specific node including conditions, addresses, and resource usage

Get logs from a Kubernetes node (kubelet, kube-proxy, or other system service logs)

Check kubelet health and recent logs across all cluster nodes. Reports service status and error/warning counts.

Check CRI-O container runtime health and recent logs across all cluster nodes.

Get detailed resource usage statistics from a node via kubelet's Summary API

Send a custom notification to Slack, Teams, or PagerDuty

Forward currently firing Alertmanager alerts to a notification channel

Send the latest scheduled cluster health report to a notification channel

Send a test notification to verify a channel is correctly configured

Resolve a PagerDuty incident by dedup key

Show which notification channels are configured (does not send anything)

Check health of all installed operators: CSV phase, subscription status, InstallPlan approval, and CatalogSource availability

Check which operators have pending upgrades and their approval strategy

Get service mesh topology graph showing namespaces, health, and traffic flow

List or get Istio configuration objects (VirtualService, DestinationRule, Gateway, etc.)

Create, patch, or delete Istio configuration objects

Get details or list services/workloads within the service mesh