OpenShift AI MCP Server with CI/CD Automation

A comprehensive Model Context Protocol (MCP) server that runs on OpenShift AI, providing both ML inference capabilities and automated CI/CD pipelines for Kubernetes/OpenShift deployments.

🚀 Features

ML Inference

Multi-Model Support: PyTorch, transformers, scikit-learn
Text & Numeric Processing: Handle both text embeddings and numeric data
Auto-Scaling: Horizontal pod autoscaling based on load
Model Registry: Dynamic model loading and management

CI/CD Automation

Git Repository Monitoring: Automatic detection of new commits
Container Image Building: Docker and OpenShift BuildConfigs
Registry Management: Push to multiple container registries
Automated Deployment: Zero-downtime deployments to OpenShift
Pipeline Management: End-to-end CI/CD pipeline orchestration

MCP Integration

Native Kubernetes API: Direct interaction without external tools
OpenShift Extensions: Routes, BuildConfigs, ImageStreams
Real-time Monitoring: Live status updates and event streaming
Security: RBAC integration and secure credential management

📋 Prerequisites

OpenShift 4.10+ or Kubernetes 1.21+
Container registry access (Quay.io, Docker Hub, etc.)
Git repository access
OpenShift AI or similar ML platform (optional but recommended)

🛠️ Installation

Quick Start: Deploy to OpenShift

# Clone the repository
git clone https://github.com/sur309/openshift-mcp-server.git
cd openshift-mcp-server

# Create namespace and apply manifests
oc apply -f manifests/namespace.yaml
oc apply -f manifests/rbac.yaml
oc apply -f manifests/configmap.yaml
oc apply -f manifests/secrets.yaml
oc apply -f manifests/deployment.yaml
oc apply -f manifests/service.yaml

# Apply Security Context Constraints (SCC) for the service account
oc adm policy add-scc-to-user anyuid -z openshift-ai-mcp-server -n ai-mcp-openshift

🚀 OpenShift Deployment Guide

Prerequisites for OpenShift

OpenShift 4.10+ cluster access
oc CLI tool installed and authenticated
Cluster admin privileges for SCC management
Container registry credentials (Quay.io, Docker Hub, etc.)

Step-by-Step OpenShift Deployment

1. Clone and Navigate to Repository

git clone https://github.com/sur309/openshift-mcp-server.git
cd openshift-mcp-server

2. Create Project/Namespace

# Create the project namespace
oc apply -f manifests/namespace.yaml

# Alternatively, create using oc new-project
# oc new-project ai-mcp-openshift --display-name="AI MCP OpenShift Server"

3. Apply Security Context Constraints (Critical for OpenShift)

OpenShift requires specific Security Context Constraints (SCC) to allow the MCP server to run with the necessary permissions:

# Add anyuid SCC to the service account
oc adm policy add-scc-to-user anyuid -z openshift-ai-mcp-server -n ai-mcp-openshift

# Verify the SCC assignment
oc describe scc anyuid | grep -A 10 "Users:"
oc get scc anyuid -o yaml | grep -A 20 users:

4. Configure RBAC (Role-Based Access Control)

# Apply service account and cluster-wide permissions
oc apply -f manifests/rbac.yaml

# Verify RBAC configuration
oc get serviceaccount openshift-ai-mcp-server -n ai-mcp-openshift
oc get clusterrolebinding openshift-ai-mcp-server

5. Create Configuration and Secrets

# Apply configuration
oc apply -f manifests/configmap.yaml

# Create registry credentials secret
oc create secret generic registry-credentials \
  --from-literal=username=YOUR_REGISTRY_USERNAME \
  --from-literal=password=YOUR_REGISTRY_TOKEN \
  --from-literal=email=YOUR_EMAIL \
  -n ai-mcp-openshift

# Create Git credentials secret
oc create secret generic git-credentials \
  --from-literal=username=YOUR_GIT_USERNAME \
  --from-literal=token=YOUR_GIT_TOKEN \
  -n ai-mcp-openshift

# Create webhook secret for CI/CD
oc create secret generic webhook-secret \
  --from-literal=secret=$(openssl rand -hex 32) \
  -n ai-mcp-openshift

# Apply the secrets manifest (if using file-based secrets)
# oc apply -f manifests/secrets.yaml

6. Deploy the Application

# Deploy the MCP server
oc apply -f manifests/deployment.yaml

# Create the service
oc apply -f manifests/service.yaml

# Verify deployment status
oc get pods -n ai-mcp-openshift
oc get deployment openshift-ai-mcp-server -n ai-mcp-openshift

7. Create OpenShift Route (for external access)

# Create route for external access
oc expose service openshift-ai-mcp-server --port=8080 -n ai-mcp-openshift

# Create route for MCP endpoint
oc create route edge openshift-ai-mcp-server-mcp \
  --service=openshift-ai-mcp-server \
  --port=8081 \
  -n ai-mcp-openshift

# Get route URLs
oc get routes -n ai-mcp-openshift

8. Verify Deployment

# Check pod status
oc get pods -n ai-mcp-openshift -o wide

# View application logs
oc logs -f deployment/openshift-ai-mcp-server -n ai-mcp-openshift

# Test health endpoints
ROUTE_URL=$(oc get route openshift-ai-mcp-server -o jsonpath='{.spec.host}' -n ai-mcp-openshift)
curl -k https://$ROUTE_URL/health
curl -k https://$ROUTE_URL:8081/health/mcp

OpenShift-Specific Features

Security Context Constraints (SCC) Explained

The MCP server requires the anyuid SCC because it:

Needs to run container builds and manage container registries
Requires access to Docker/Podman socket for CI/CD operations
May need to run with specific user IDs for compatibility

# Check current SCC assignments
oc get scc anyuid -o yaml

# Verify service account has proper SCC
oc describe serviceaccount openshift-ai-mcp-server -n ai-mcp-openshift

OpenShift Routes vs Services

# Internal access (within cluster)
oc get svc openshift-ai-mcp-server -n ai-mcp-openshift

# External access (internet-facing)
oc get routes -n ai-mcp-openshift

# Test internal service
oc port-forward svc/openshift-ai-mcp-server 8080:8080 -n ai-mcp-openshift

Troubleshooting OpenShift Deployment

Common OpenShift Issues

SCC Permission Denied

# Error: pods "openshift-ai-mcp-server-xxx" is forbidden: unable to validate against any security context constraint
oc adm policy add-scc-to-user anyuid -z openshift-ai-mcp-server -n ai-mcp-openshift

Image Pull Issues

# Check image pull secrets
oc get secrets -n ai-mcp-openshift | grep registry
oc describe pod <pod-name> -n ai-mcp-openshift

RBAC Permission Errors

# Check cluster role binding
oc get clusterrolebinding openshift-ai-mcp-server -o yaml
oc auth can-i create pods --as=system:serviceaccount:ai-mcp-openshift:openshift-ai-mcp-server

Route Access Issues

# Check route configuration
oc describe route openshift-ai-mcp-server -n ai-mcp-openshift

# Test internal connectivity
oc rsh deployment/openshift-ai-mcp-server curl localhost:8080/health

Debug Commands

# Get comprehensive pod information
oc describe pod -l app.kubernetes.io/name=openshift-ai-mcp-server -n ai-mcp-openshift

# Check resource usage
oc adm top pod -n ai-mcp-openshift

# View events
oc get events --sort-by='.lastTimestamp' -n ai-mcp-openshift

# Check security context
oc get pod -o yaml | grep -A 10 securityContext

Alternative Installation Methods

1. Using oc new-app (One-Command Deployment)

# Create new project and deploy in one command
oc new-project ai-mcp-openshift
oc new-app https://github.com/sur309/openshift-mcp-server.git \
  --name=openshift-ai-mcp-server \
  --strategy=docker

# Apply SCC and additional configurations
oc adm policy add-scc-to-user anyuid -z default -n ai-mcp-openshift
oc expose svc/openshift-ai-mcp-server

2. Using Helm Charts (if available)

# Add Helm repository (if published)
helm repo add openshift-mcp-server https://your-helm-repo.com
helm repo update

# Install with Helm
helm install mcp-server openshift-mcp-server/openshift-mcp-server \
  --namespace ai-mcp-openshift \
  --create-namespace \
  --set image.tag=latest \
  --set serviceAccount.annotations."openshift\.io/scc"=anyuid

🔧 Configuration

Container Image Management

Build and Push Container Image

# Build the container image
docker build -t quay.io/YOUR_ORG/ai-mcp-openshift-server:latest .

# Push to registry
docker push quay.io/YOUR_ORG/ai-mcp-openshift-server:latest

# Update deployment image
oc set image deployment/ai-mcp-openshift-server \
  inference-server=quay.io/YOUR_ORG/ai-mcp-openshift-server:latest \
  -n ai-mcp-openshift

🔧 Configuration

Environment Variables

Variable	Description	Default
`PORT`	Inference server port	`8080`
`MCP_PORT`	MCP server port	`8081`
`MCP_PROFILE`	MCP profile to use	`cicd`
`LOG_LEVEL`	Logging verbosity (0-9)	`2`
`DEFAULT_REGISTRY`	Default container registry	`quay.io`
`DEFAULT_NAMESPACE`	Default deployment namespace	`ai-mcp-openshift`
`MODELS_PATH`	Path to ML models	`/app/models`

MCP Profiles

full: All tools including CI/CD, Kubernetes management, and Helm
cicd: Focused on CI/CD operations with essential Kubernetes tools

📖 Usage Examples

1. Set up a CI/CD Pipeline

# Add a container registry
curl -X POST http://localhost:8081/mcp \
  -H "Content-Type: application/json" \
  -d '{
    "tool": "registry_add",
    "arguments": {
      "name": "quay",
      "url": "quay.io",
      "username": "your-username",
      "password": "your-token"
    }
  }'

# Create a complete CI/CD pipeline
curl -X POST http://localhost:8081/mcp \
  -H "Content-Type: application/json" \
  -d '{
    "tool": "cicd_create_pipeline",
    "arguments": {
      "name": "my-app-pipeline",
      "git_url": "https://github.com/user/my-app.git",
      "git_branch": "main",
      "image_name": "my-app",
      "registry": "quay",
      "deploy_namespace": "my-app-prod",
      "dockerfile": "Dockerfile",
      "env_vars": {
        "NODE_ENV": "production",
        "PORT": "3000"
      }
    }
  }'

2. Manual Build and Deploy

# Build an image
curl -X POST http://localhost:8081/mcp \
  -H "Content-Type: application/json" \
  -d '{
    "tool": "build_image",
    "arguments": {
      "name": "my-build",
      "source_repo": "https://github.com/user/my-app.git",
      "image_name": "my-app",
      "image_tag": "v1.0.0"
    }
  }'

# Push to registry
curl -X POST http://localhost:8081/mcp \
  -H "Content-Type: application/json" \
  -d '{
    "tool": "registry_push",
    "arguments": {
      "source_image": "my-app:v1.0.0",
      "target_image": "quay.io/user/my-app",
      "target_tag": "v1.0.0",
      "registry": "quay"
    }
  }'

# Deploy application
curl -X POST http://localhost:8081/mcp \
  -H "Content-Type: application/json" \
  -d '{
    "tool": "deploy_application",
    "arguments": {
      "name": "my-app",
      "image": "quay.io/user/my-app",
      "tag": "v1.0.0",
      "namespace": "my-app-prod",
      "replicas": 3,
      "port": 3000,
      "expose_route": true
    }
  }'

3. ML Inference

# Text embedding inference
curl -X POST http://localhost:8080/infer \
  -H "Content-Type: application/json" \
  -d '{
    "inputs": ["Hello world", "OpenShift AI is great"],
    "model_name": "text_embeddings"
  }'

# Numeric inference
curl -X POST http://localhost:8080/infer \
  -H "Content-Type: application/json" \
  -d '{
    "inputs": [[1.0, 2.0, 3.0, 4.0]],
    "model_name": "simple_classifier"
  }'

# List available models
curl http://localhost:8080/models

✅ Post-Deployment Verification

Verify MCP Server Installation

After deploying the MCP server to OpenShift, run these commands to ensure everything is working correctly:

# 1. Check pod status
oc get pods -l app.kubernetes.io/name=openshift-ai-mcp-server -n ai-mcp-openshift

# 2. Verify service account has proper SCC
oc get serviceaccount openshift-ai-mcp-server -n ai-mcp-openshift -o yaml

# 3. Check SCC assignment
oc get scc anyuid -o yaml | grep -A 10 "users:"

# 4. Test health endpoints
ROUTE_URL=$(oc get route openshift-ai-mcp-server -o jsonpath='{.spec.host}' -n ai-mcp-openshift 2>/dev/null)
if [ ! -z "$ROUTE_URL" ]; then
  echo "Testing health endpoint: https://$ROUTE_URL/health"
  curl -k -s https://$ROUTE_URL/health | jq .
else
  echo "Route not found. Testing via port-forward..."
  oc port-forward svc/openshift-ai-mcp-server 8080:8080 -n ai-mcp-openshift &
  sleep 2
  curl -s http://localhost:8080/health | jq .
  pkill -f "oc port-forward"
fi

# 5. Test MCP endpoint
curl -k -s https://$ROUTE_URL:8081/mcp/tools | jq .

# 6. Check application logs
oc logs -f deployment/openshift-ai-mcp-server --tail=50 -n ai-mcp-openshift

Using the MCP CLI Tool

After deployment, you can use the included CLI tool to interact with your MCP server:

# Navigate to the CLI tool directory
cd cursor-integration/cli-tools

# Set the MCP server URL (replace with your actual route)
export MCP_SERVER_URL="https://$(oc get route openshift-ai-mcp-server-mcp -o jsonpath='{.spec.host}' -n ai-mcp-openshift)"

# Test the CLI tool
node mcp-cli.js tools

# Execute a workflow
node mcp-cli.js execute "Deploy a sample application to my OpenShift cluster"

# Analyze a prompt
node mcp-cli.js analyze "I want to build and deploy a container from my Git repository"

🔄 CI/CD Pipeline Flow

Git Monitoring: Server monitors specified Git repositories for new commits
Automatic Trigger: New commits trigger the CI/CD pipeline
Image Build: Source code is built into a container image
Registry Push: Built image is pushed to the configured registry
Deployment: Application is deployed to the specified OpenShift project
Verification: Health checks ensure successful deployment

🏗️ Architecture

┌─────────────────┐    ┌─────────────────┐    ┌─────────────────┐
│   Git Repo      │    │   MCP Server    │    │   Inference     │
│                 │    │                 │    │   Engine        │
│ ┌─────────────┐ │    │ ┌─────────────┐ │    │ ┌─────────────┐ │
│ │   Commit    │ │───▶│ │  Git Watch  │ │    │ │  ML Models  │ │
│ │   Events    │ │    │ │             │ │    │ │             │ │
│ └─────────────┘ │    │ └─────────────┘ │    │ └─────────────┘ │
└─────────────────┘    │ ┌─────────────┐ │    │ ┌─────────────┐ │
                       │ │Image Builder│ │    │ │  FastAPI    │ │
┌─────────────────┐    │ │             │ │    │ │  Server     │ │
│Container Registry│◀───│ └─────────────┘ │    │ └─────────────┘ │
│                 │    │ ┌─────────────┐ │    └─────────────────┘
│ ┌─────────────┐ │    │ │  Deploy     │ │
│ │   Images    │ │    │ │  Manager    │ │    ┌─────────────────┐
│ │             │ │    │ │             │ │    │   OpenShift     │
│ └─────────────┘ │    │ └─────────────┘ │───▶│   Cluster       │
└─────────────────┘    └─────────────────┘    │                 │
                                              │ ┌─────────────┐ │
                                              │ │   Pods      │ │
                                              │ │   Services  │ │
                                              │ │   Routes    │ │
                                              │ └─────────────┘ │
                                              └─────────────────┘

🔧 Troubleshooting

Common Issues

Build Failures

# Check build logs
oc logs deployment/ai-mcp-openshift-server -n ai-mcp-openshift

Registry Access Issues

# Verify registry credentials
oc get secret registry-credentials -o yaml -n ai-mcp-openshift

Git Authentication

# Check Git credentials
oc get secret git-credentials -o yaml -n ai-mcp-openshift

Debug Commands

# Check server status
curl http://localhost:8080/health
curl http://localhost:8081/health/mcp

# List all MCP tools
curl http://localhost:8081/mcp/tools

# View server logs
oc logs -f deployment/ai-mcp-openshift-server -n ai-mcp-openshift

# Check resource usage
oc top pods -n ai-mcp-openshift

🤝 Contributing

Fork the repository
Create a feature branch (git checkout -b feature/amazing-feature)
Commit your changes (git commit -m 'Add amazing feature')
Push to the branch (git push origin feature/amazing-feature)
Open a Pull Request

📄 License

This project is licensed under the MIT License - see the LICENSE file for details.

🙏 Acknowledgments

Model Context Protocol (MCP) for the protocol specification
OpenShift for the container platform
Kubernetes for the orchestration platform

Name		Name	Last commit message	Last commit date
Latest commit History 278 Commits
.github		.github
cmd		cmd
cursor-integration		cursor-integration
docs/images		docs/images
manifests		manifests
npm		npm
pkg		pkg
python		python
.gitignore		.gitignore
CLAUDE_INTEGRATION.md		CLAUDE_INTEGRATION.md
COPILOT_INTEGRATION.md		COPILOT_INTEGRATION.md
DEMO_COMPLETE_SUCCESS.md		DEMO_COMPLETE_SUCCESS.md
DEMO_SAMPLE_GAMING_APP.md		DEMO_SAMPLE_GAMING_APP.md
DEPLOYMENT.md		DEPLOYMENT.md
Dockerfile		Dockerfile
Dockerfile.build-macos		Dockerfile.build-macos
Dockerfile.build-new		Dockerfile.build-new
Dockerfile.container-runtime		Dockerfile.container-runtime
Dockerfile.minimal		Dockerfile.minimal
Dockerfile.openshift		Dockerfile.openshift
Dockerfile.optimized		Dockerfile.optimized
Dockerfile.privileged		Dockerfile.privileged
FINAL_FIX_SUMMARY.md		FINAL_FIX_SUMMARY.md
INTELLIGENT_CONTAINER_WORKFLOWS.md		INTELLIGENT_CONTAINER_WORKFLOWS.md
LICENSE		LICENSE
MCP_BEST_PRACTICES_COMPLIANCE.md		MCP_BEST_PRACTICES_COMPLIANCE.md
Makefile		Makefile
OPTIMIZATION.md		OPTIMIZATION.md
QUICKSTART.md		QUICKSTART.md
README.md		README.md
README_OLD.md		README_OLD.md
TROUBLESHOOTING.md		TROUBLESHOOTING.md
VSCODE_INTEGRATION.md		VSCODE_INTEGRATION.md
deploy.sh		deploy.sh
fix-openshift-scc.sh		fix-openshift-scc.sh
fix-security.sh		fix-security.sh
gaming-app-manifests.yaml		gaming-app-manifests.yaml
go.mod		go.mod
go.sum		go.sum
mcp-http-bridge.js		mcp-http-bridge.js
openshift-mcp-server		openshift-mcp-server
openshift-mcp-server-new		openshift-mcp-server-new
openshift-mcp-server-old		openshift-mcp-server-old
package-lock.json		package-lock.json
package.json		package.json
simple-mcp-bridge.js		simple-mcp-bridge.js
smithery.yaml		smithery.yaml
test-bridge.js		test-bridge.js
test-exact-flow.js		test-exact-flow.js
test-tools-list.js		test-tools-list.js
validate-manifests.sh		validate-manifests.sh

Folders and files

Latest commit

History

Repository files navigation

OpenShift AI MCP Server with CI/CD Automation

🚀 Features

ML Inference

CI/CD Automation

MCP Integration

📋 Prerequisites

🛠️ Installation

Quick Start: Deploy to OpenShift

🚀 OpenShift Deployment Guide

Prerequisites for OpenShift

Step-by-Step OpenShift Deployment

1. Clone and Navigate to Repository

2. Create Project/Namespace

3. Apply Security Context Constraints (Critical for OpenShift)

4. Configure RBAC (Role-Based Access Control)

5. Create Configuration and Secrets

6. Deploy the Application

7. Create OpenShift Route (for external access)

8. Verify Deployment

OpenShift-Specific Features

Security Context Constraints (SCC) Explained

OpenShift Routes vs Services

Troubleshooting OpenShift Deployment

Common OpenShift Issues

Debug Commands

Alternative Installation Methods

1. Using oc new-app (One-Command Deployment)

2. Using Helm Charts (if available)

🔧 Configuration

Container Image Management

Build and Push Container Image

🔧 Configuration

Environment Variables

MCP Profiles

📖 Usage Examples

1. Set up a CI/CD Pipeline

2. Manual Build and Deploy

3. ML Inference

✅ Post-Deployment Verification

Verify MCP Server Installation

Using the MCP CLI Tool

🔄 CI/CD Pipeline Flow

🏗️ Architecture

🔧 Troubleshooting

Common Issues

Debug Commands

🤝 Contributing

📄 License

🙏 Acknowledgments

About

Resources

License

Uh oh!

Stars

Watchers

Forks

Releases

Packages 0

Uh oh!

Contributors

Uh oh!

Languages

Packages