Deployment

Deploy EVOID applications to production.

Quick Deploy

evo service run api

This starts the server at http://0.0.0.0:8000.

Production Setup

Uvicorn with Workers

uvicorn main:app --host 0.0.0.0 --port 8000 --workers 4

Worker count: CPU cores * 2 + 1

Systemd Service

[Unit]
Description=EVOID API Service
After=network.target

[Service]
Type=exec
User=www-data
WorkingDirectory=/opt/my-api
ExecStart=/opt/my-api/.venv/bin/evo service run api
Restart=always

[Install]
WantedBy=multi-user.target

sudo systemctl enable my-api
sudo systemctl start my-api

Docker

FROM python:3.13-slim
WORKDIR /app
COPY pyproject.toml .
RUN pip install uv && uv sync
COPY . .
EXPOSE 8000
CMD ["evo", "service", "run", "api"]

docker build -t my-api .
docker run -p 8000:8000 my-api

Docker Compose

services:
  api:
    build: .
    ports:
      - "8000:8000"
    environment:
      - DATABASE_URL=postgresql://...
    restart: unless-stopped

SSL / HTTPS

Nginx Reverse Proxy

server {
    listen 443 ssl;
    server_name myapi.example.com;

    ssl_certificate /etc/letsencrypt/live/myapi.example.com/fullchain.pem;
    ssl_certificate_key /etc/letsencrypt/live/myapi.example.com/privkey.pem;

    location / {
        proxy_pass http://127.0.0.1:8000;
        proxy_set_header Host $host;
        proxy_set_header X-Real-IP $remote_addr;
    }
}

Health Check

from evoid.adapters.asgi import get
from evoid.web.route import Service

app = Service("my-api")

@get("/health", level="ephemeral")
async def health() -> dict:
    return {"status": "healthy"}

Environment Variables

import os

DATABASE_URL = os.getenv("DATABASE_URL", "sqlite:///db.sqlite")
REDIS_URL = os.getenv("REDIS_URL", "redis://localhost")
DEBUG = os.getenv("DEBUG", "false").lower() == "true"

Graceful Shutdown

import signal
import sys

def shutdown(sig, frame):
    # Cleanup resources
    sys.exit(0)

signal.signal(signal.SIGTERM, shutdown)
signal.signal(signal.SIGINT, shutdown)

Monitoring

Process Management

systemctl status my-api
journalctl -u my-api -f
systemctl restart my-api

Metrics

from evoid.engines.metrics import simple as metrics

metrics.increment("requests.total")
start = metrics.timer_start("request.duration")
# ... handle request
metrics.timer_stop("request.duration", start)

Best Practices

• Use level="ephemeral" for health check endpoints
• Set appropriate timeouts per Intent level
• Use structured logging in production
• Monitor pipeline execution times
• Use the message bus for inter-service communication, not HTTP