AI-Powered OpenShift Assistant

OpenShift Lightspeed

AI-powered assistant integrated directly into the OpenShift console. Ask questions in natural language, get instant answers about your cluster, troubleshoot issues, and learn OpenShift concepts — all without leaving the console.

Lightspeed + MCP in Action

Demo: Using OpenShift Lightspeed with MCP (Model Context Protocol) integration on a live cluster

💡 What is OpenShift Lightspeed?

OpenShift Lightspeed is Red Hat's AI-powered assistant built into the OpenShift web console. It uses large language models (LLMs) to help administrators and developers interact with their clusters using natural language.

Instead of searching documentation or memorizing CLI commands, users can simply ask questions like "Why is my pod crashing?" or "How do I set up a NetworkPolicy?" and get contextual, actionable answers.

Key Insight: Lightspeed is not a generic chatbot — it has context about your specific OpenShift cluster, including resources, events, and configurations.

🚀 Key Features

🤖

Natural Language Queries

Ask questions in plain English. No need to memorize oc commands or YAML syntax — Lightspeed translates intent to actions.

🔍

Cluster-Aware Context

Lightspeed understands your cluster state — pods, deployments, events, logs — and provides answers specific to your environment.

🛠️

Troubleshooting Assistant

Diagnose pod failures, resource issues, and misconfigurations. Get step-by-step remediation guidance.

📚

Learning & Documentation

Learn OpenShift concepts inline. Lightspeed pulls from Red Hat documentation to explain features and best practices.

🔌

MCP Integration

Model Context Protocol enables Lightspeed to interact with external tools, APIs, and data sources for richer responses.

🔒

Enterprise Security

All queries are processed through Red Hat's secure infrastructure. No cluster data leaves your control boundary without consent.

🏗️ Architecture Overview

Lightspeed operates as an Operator-managed service within the OpenShift cluster, communicating with an LLM backend via Red Hat's AI infrastructure.

Component	Role	Location
Lightspeed UI Plugin	Console plugin providing chat interface	OpenShift Console
OLS API Server	Backend service processing queries, managing conversation	In-cluster (Operator)
RAG Engine	Retrieval-Augmented Generation from Red Hat docs	In-cluster
MCP Server	Model Context Protocol for tool/API integrations	In-cluster
LLM Provider	Large language model (Granite, GPT-4, etc.)	Red Hat AI / External
Guardrails	Content filtering, prompt safety, response validation	AI Backend

🔌 MCP (Model Context Protocol) Integration

MCP extends Lightspeed beyond simple Q&A by allowing the LLM to invoke tools and access live cluster data. This enables actions like querying pod status, reading logs, or checking resource utilization — all through natural language.

What MCP Enables: Instead of just answering from documentation, Lightspeed can actively inspect your cluster state, read real-time metrics, and provide answers based on actual data.

Example MCP Capabilities

Cluster inspection — Query pod status, deployment rollouts, node health
Log analysis — Read and interpret pod/container logs
Resource auditing — Check quota usage, resource requests/limits
Configuration review — Inspect ConfigMaps, Secrets (metadata), Routes
Event correlation — Connect Kubernetes events to root causes

How It Works

User: "Why is my nginx pod in CrashLoopBackOff?"

Lightspeed (via MCP):
  1. oc get pod nginx-xyz -o yaml     → checks pod spec
  2. oc logs nginx-xyz --previous      → reads crash logs
  3. oc describe pod nginx-xyz         → checks events

Answer: "Your nginx pod is crashing because the ConfigMap
'nginx-config' referenced in the volume mount doesn't exist
in namespace 'web-app'. Create the ConfigMap or update the
deployment to remove the volume mount."

⚙️ Installation

Lightspeed is deployed via the OpenShift Lightspeed Operator from OperatorHub.

Prerequisites

OpenShift 4.15+ cluster
Cluster-admin access
LLM provider credentials (Red Hat AI, Azure OpenAI, or self-hosted)

Quick Install Steps

# 1. Install the Operator from OperatorHub
#    Navigate to: Operators → OperatorHub → Search "Lightspeed"

# 2. Create LLM provider secret
oc create secret generic llm-credentials \
  --from-literal=apitoken='your-api-key' \
  -n openshift-lightspeed

# 3. Create OLSConfig custom resource
cat <<EOF | oc apply -f -
apiVersion: ols.openshift.io/v1alpha1
kind: OLSConfig
metadata:
  name: cluster
spec:
  llm:
    providers:
      - name: my-provider
        type: openai
        credentialsSecretRef:
          name: llm-credentials
        url: "https://api.openai.com/v1"
        models:
          - name: gpt-4
  ols:
    defaultModel: gpt-4
    defaultProvider: my-provider
    logLevel: INFO
EOF

Note: For production, Red Hat recommends using IBM Granite models through Red Hat AI (RHAI) or RHEL AI for on-premise deployments with full data sovereignty.

🎯 Common Use Cases

Scenario	Example Query	What Lightspeed Does
Troubleshooting	"Why is my deployment failing?"	Checks events, pod status, resource limits, image pull errors
Learning	"What is a PersistentVolumeClaim?"	Explains concept with Red Hat docs context
Configuration	"How do I add a health check to my deployment?"	Provides YAML with readiness/liveness probes
Security	"Which pods are running as root?"	Queries cluster via MCP, lists privileged pods
Networking	"Create a NetworkPolicy to isolate namespace X"	Generates NetworkPolicy YAML with explanation
Scaling	"What's the CPU usage of my app?"	Reads metrics, suggests HPA configuration

⚠️ Unofficial Document — This content is not affiliated with, endorsed by, or officially associated with Red Hat, Inc. or IBM. It is created for educational and reference purposes only. OpenShift, Red Hat, and related trademarks are the property of Red Hat, Inc.