PM2 process management: ecosystem config and zero-downtime reload

Running node server.js directly means the process dies on crash, doesn't restart on server reboot, and restarts cause brief downtime.

Use pm2 with an ecosystem config file:

// ecosystem.config.js
module.exports = {
  apps: [{
    name: 'myapp',
    script: './dist/server.js',
    instances: 'max',       // Use all CPU cores
    exec_mode: 'cluster',   // Required for multiple instances
    watch: false,
    max_memory_restart: '1G',
    env: {
      NODE_ENV: 'development',
    },
    env_production: {
      NODE_ENV: 'production',
      PORT: 3000,
    },
    error_file: '/var/log/myapp/error.log',
    out_file: '/var/log/myapp/out.log',
    log_date_format: 'YYYY-MM-DD HH:mm:ss',
  }]
};

# Start / reload
pm2 start ecosystem.config.js --env production
pm2 reload myapp          # Zero-downtime reload
pm2 save                  # Save process list
pm2 startup               # Generate startup script

Cluster mode distributes connections across worker processes. pm2 reload performs rolling restarts — one worker at a time — so there is no downtime during deploys.