📚 Documentation Issue Summary

The documentation currently lacks a guide for deploying MCP Gateway to [Google Cloud Run](https://cloud.google.com/run), a fully managed serverless platform for containers with automatic HTTPS, scale-to-zero, and built-in support for Cloud SQL and Memorystore (Redis).

📍 Location of the Problem

Suggested location: docs/docs/deployment/google-cloud-run.md

✏️ Type of Issue

• Missing explanation or example
• Unclear deployment coverage

💡 Suggested Fix or Clarification

Create a new deployment guide following the format used in docs/docs/deployment/ibm-code-engine.md. Include end-to-end setup for Cloud SQL (Postgres), Memorystore (Redis), JWT auth, and curl-based health checks.

✅ Overview

• Explain Cloud Run as a secure, auto-scaling container platform.
• Emphasize native support for HTTPS, public endpoints, and scaling down to zero.
• Note that the ghcr.io/ibm/mcp-context-forge:latest image can be used directly—no build required.

🛠 Prerequisites

• Google Cloud account + billing enabled

• gcloud CLI installed and initialized

• Required services enabled:

  • run.googleapis.com
  • sqladmin.googleapis.com
  • redis.googleapis.com

• PostgreSQL and Redis instances created

• .env values ready (JWT_SECRET_KEY, DATABASE_URL, REDIS_URL)

⚙️ Setup Steps

1. Provision Cloud SQL (Postgres)

gcloud sql instances create mcpgw-db \
  --database-version=POSTGRES_14 \
  --cpu=2 --memory=4GiB --region=us-central1

gcloud sql users set-password postgres \
  --instance=mcpgw-db --password=mysecretpassword

gcloud sql databases create mcpgw --instance=mcpgw-db

2. Provision Memorystore (Redis)

gcloud redis instances create mcpgw-redis \
  --region=us-central1 \
  --tier=STANDARD_HA \
  --size=1

3. Deploy to Cloud Run

gcloud run deploy mcpgateway \
  --image=ghcr.io/ibm/mcp-context-forge:latest \
  --region=us-central1 \
  --platform=managed \
  --allow-unauthenticated \
  --port=4444 \
  --cpu=2 --memory=2Gi \
  --set-env-vars=\
JWT_SECRET_KEY=your-secret,\
BASIC_AUTH_USER=admin,\
BASIC_AUTH_PASSWORD=changeme,\
AUTH_REQUIRED=true,\
DATABASE_URL=postgresql://postgres:mysecretpassword@<SQL_IP>:5432/mcpgw,\
REDIS_URL=redis://<REDIS_IP>:6379/0,\
CACHE_TYPE=redis

🔒 Auth & Access

Generate a JWT token:

docker run -it --rm ghcr.io/ibm/mcp-context-forge:latest \
  python3 -m mcpgateway.utils.create_jwt_token -u admin

Call a protected endpoint:

curl -H "Authorization: Bearer $MCPGATEWAY_BEARER_TOKEN" \
     https://<your-cloud-run-url>/tools

📝 Features to Document

• Native HTTPS with no TLS config
• Environment injection via --set-env-vars
• JWT generation via CLI or container run
• Postgres + Redis integration via GCP IPs
• Optional: build automation via GitHub Actions (deploy-gcr.yml)

🧩 Additional Notes

• Use gcloud sql instances describe and gcloud redis instances describe to get IPs
• Align structure and tone with ibm-code-engine.md for consistency
• Cloud Run supports public/private endpoints, autoscaling, and container revision history
• Consider linking to cloudbuild.yaml or GitHub CI workflows if build automation is added later