Gunicorn¶

Industry-standard WSGI server for Django. Battle-tested and widely deployed.

Use when: Traditional Django apps on Linux Don’t use when: Windows (use Waitress), async/WebSockets (use ASGI Mode)

Installation¶

pip install django-prodserver[gunicorn]

Configuration¶

PRODUCTION_PROCESSES = {
    "web": {
        "BACKEND": "django_prodserver.backends.servers.gunicorn.GunicornServer",
        "ARGS": {
            "bind": "0.0.0.0:8000",
            "workers": "4",
            "timeout": "60",
        }
    }
}

Common ARGS¶

Argument	Default	Description
`bind`	`127.0.0.1:8000`	Socket to bind (host:port or unix socket)
`workers`	`1`	Worker processes (recommend: `2-4 x CPU`)
`worker-class`	`sync`	Worker type (`sync`, `gevent`, `eventlet`)
`timeout`	`30`	Request timeout (seconds)
`max-requests`	`0`	Restart workers after N requests
`max-requests-jitter`	`0`	Random jitter for max-requests
`keepalive`	`2`	Keep-alive timeout
`accesslog`	`None`	Access log path (`-` for stdout)
`errorlog`	`stderr`	Error log path
`loglevel`	`info`	Log level

Examples¶

Production¶

"ARGS": {
    "bind": "0.0.0.0:8000",
    "workers": "8",
    "worker-class": "gevent",
    "timeout": "120",
    "max-requests": "1000",
    "max-requests-jitter": "100",
}

Unix Socket (behind nginx)¶

"ARGS": {
    "bind": "unix:/var/run/myapp.sock",
    "workers": "4",
    "user": "www-data",
}

Gevent (I/O-bound apps)¶

pip install gunicorn[gevent]

"ARGS": {
    "worker-class": "gevent",
    "worker-connections": "1000",
    "workers": "4",
}

Worker Count¶

Formula: (2 x CPU_cores) + 1

CPU-bound: Fewer workers (1-2 per core)
I/O-bound: More workers or use gevent

Troubleshooting¶

Worker timeout: Increase timeout or investigate slow requests

High memory: Reduce workers, add max-requests for recycling

Port in use: Change port or stop conflicting process