Skip to content

samples

Directory actions

More options

Directory actions

More options

Latest commit

 

History

History

samples

README.md

MCPServer Samples

Example configurations for deploying MCP servers with the operator.

Quick Reference

# Sample Transport Metrics Use Case
01 wikipedia-sse SSE No Minimal example using Wikipedia MCP
02 streamable-http-basic Streamable HTTP No Modern transport, production-ready
03 sse-optimized SSE No SSE with session affinity
04 metrics-basic Streamable HTTP Yes Basic metrics collection
05 metrics-advanced Streamable HTTP Yes Custom sidecar configuration
06 metrics-sse SSE Yes SSE with metrics
10 complete-reference Auto Yes All available options

Decision Tree

Which sample should I use?
│
├── Just getting started?
│   └── 01-wikipedia-sse.yaml (minimal, works out of the box)
│
├── Building a new server?
│   └── 02-streamable-http-basic.yaml (modern transport)
│
├── Using legacy SSE server?
│   ├── Simple deployment → 01-wikipedia-sse.yaml
│   └── Need optimizations → 03-sse-optimized.yaml
│
├── Need metrics?
│   ├── Streamable HTTP → 04-metrics-basic.yaml
│   ├── Custom config → 05-metrics-advanced.yaml
│   └── SSE + metrics → 06-metrics-sse.yaml
│
└── Reference for all options?
    └── 10-complete-reference.yaml

Sample Descriptions

01-wikipedia-sse.yaml

Minimal example using the public Wikipedia MCP server image. Great for:

  • Quick testing
  • Learning the basics
  • Verifying operator installation
kubectl apply -f 01-wikipedia-sse.yaml
kubectl port-forward svc/wikipedia-mcp 3001:3001
# Test: curl -N http://localhost:3001/sse

02-streamable-http-basic.yaml

Production-ready configuration using modern Streamable HTTP transport:

  • Explicit protocol configuration
  • HPA enabled
  • Metrics collection
  • Health checks

03-sse-optimized.yaml

SSE configuration with optimizations for production:

  • Session affinity for sticky connections
  • Termination grace period for graceful shutdowns
  • Suitable for stateful SSE connections

04-metrics-basic.yaml

Basic metrics collection with mcp-proxy sidecar:

  • Prometheus metrics on port 9090
  • Auto-creates ServiceMonitor if Prometheus Operator is installed
  • Minimal configuration

05-metrics-advanced.yaml

Advanced metrics with custom sidecar configuration:

  • Custom resource limits for sidecar
  • TLS termination example
  • Extended configuration options

06-metrics-sse.yaml

SSE transport with metrics collection:

  • Combines SSE optimizations with metrics
  • Session affinity + metrics sidecar
  • Production SSE deployment

10-complete-reference.yaml

Reference example showing all available CRD fields:

  • Every configuration option documented
  • Not meant for direct use
  • Copy sections as needed

Usage

Apply a sample

kubectl apply -f 01-wikipedia-sse.yaml

Watch status

kubectl get mcpserver -w

Test connection

# Port forward
kubectl port-forward svc/<mcpserver-name> 8080:8080

# Test (SSE)
curl -N http://localhost:8080/sse

# Test (Streamable HTTP)
curl -X POST http://localhost:8080/mcp \
  -H "Content-Type: application/json" \
  -d '{"jsonrpc":"2.0","id":1,"method":"initialize","params":{}}'

Advanced Examples

For production patterns like NetworkPolicies and PodDisruptionBudgets, see:

Documentation