Production Checklist (Python)

Skip the checklist: Deploy to NitroCloud and it handles HTTPS, scaling, monitoring and the container build for you. See Cloud Platforms.

Pre-Deployment

Code

• All tests passing (pytest)
• Linter and type checker passing
• No stray print() calls. Use the logger, and never write to stdout if you also use stdio
• Error handling implemented
• Input validation on all tools (Pydantic models)

Security

• JWT_SECRET is strong and unique
• API keys are not hardcoded
• Secrets only in env or a secret manager, never in the wheel
• OAUTH_REQUIRED / API_KEY set if the server is not public
• Rate limiting configured
• Guards on sensitive tools
• HTTPS enabled (terminate TLS at the platform or proxy)

Configuration

• Python 3.10+ in the image
• MCP_TRANSPORT_TYPE set (http or dual)
• HOST=0.0.0.0 in containers
• PORT matches what the platform expects
• MCP_MAX_SESSIONS and session timeout set for HTTP
• Logging level set to info or warn
• CORS configured if browsers call the server

Validate and Build

nitrostack-py validate
nitrostack-py pack --dry-run   # review what goes into the wheel
nitrostack-py pack

pack never includes .env. Python has no dist/index.js build step.

Deployment

Option 1: NitroCloud (recommended)

Connect your repo or upload your project and NitroCloud builds and runs it. See Cloud Platforms.

Option 2: Plain process

HOST=0.0.0.0 MCP_TRANSPORT_TYPE=http PORT=3000 nitrostack-py start

Option 3: Docker

docker build -t my-mcp-server .
docker run --rm -p 3000:3000 --env-file .env my-mcp-server

Option 4: systemd

# /etc/systemd/system/mcp-server.service
[Unit]
Description=MCP server
After=network.target

[Service]
WorkingDirectory=/opt/mcp-server
EnvironmentFile=/etc/mcp-server.env
ExecStart=/opt/mcp-server/.venv/bin/nitrostack-py start --port 3000
Restart=on-failure
User=mcp

[Install]
WantedBy=multi-user.target

Post-Deployment

• GET /mcp/health responding
• Tools list correctly from the MCP endpoint (/mcp, no trailing slash)
• Authentication working
• Logs being written
• Error tracking configured

Monitoring

• Check logs regularly
• Monitor error rates
• Track tool usage
• Set up alerts on the health endpoint

Next Steps

• Docker Guide
• Cloud Platforms