Manage HTTP servers with CherryPy.


class cherrypy._cpserver.Server

An adapter for an HTTP server.

You can set attributes (like socket_host and socket_port) on this object (which is probably cherrypy.server), and call quickstart. For example:

cherrypy.server.socket_port = 80

Return the base (scheme://host[:port] or sock file) for this server.


A (host, port) tuple for TCP sockets or a str for Unix domain sockets.


Return a (httpserver, bind_addr) pair based on self attributes.

instance = None

If not None, this should be an HTTP server instance (such as CPWSGIServer) which cherrypy.server will control. Use this when you need more control over object instantiation than is available in the various configuration options.

max_request_body_size = 104857600

The maximum number of bytes allowable in the request body. If exceeded, the HTTP server should return “413 Request Entity Too Large”.

max_request_header_size = 512000

The maximum number of bytes allowable in the request headers. If exceeded, the HTTP server should return “413 Request Entity Too Large”.

nodelay = True

If True (the default since 3.1), sets the TCP_NODELAY socket option.

protocol_version = 'HTTP/1.1'

The version string to write in the Status-Line of all HTTP responses, for example, “HTTP/1.1” (the default). Depending on the HTTP server used, this should also limit the supported features used in the response.

shutdown_timeout = 5

The time to wait for HTTP worker threads to clean up.

socket_file = None

If given, the name of the UNIX socket to use instead of TCP/IP.

When this option is not None, the socket_host and socket_port options are ignored.


The hostname or IP address on which to listen for connections.

Host values may be any IPv4 or IPv6 address, or any valid hostname. The string ‘localhost’ is a synonym for ‘’ (or ‘::1’, if your hosts file prefers IPv6). The string ‘’ is a special IPv4 entry meaning “any active interface” (INADDR_ANY), and ‘::’ is the similar IN6ADDR_ANY for IPv6. The empty string or None are not allowed.

socket_port = 8080

The TCP port on which to listen for connections.

socket_queue_size = 5

The ‘backlog’ argument to socket.listen(); specifies the maximum number of queued connections (default 5).

socket_timeout = 10

The timeout in seconds for accepted connections (default 10).

ssl_certificate = None

The filename of the SSL certificate to use.

ssl_certificate_chain = None

When using PyOpenSSL, the certificate chain to pass to Context.load_verify_locations.

ssl_context = None

When using PyOpenSSL, an instance of SSL.Context.

ssl_module = 'pyopenssl'

The name of a registered SSL adaptation module to use with the builtin WSGI server. Builtin options are ‘builtin’ (to use the SSL library built into recent versions of Python) and ‘pyopenssl’ (to use the PyOpenSSL project, which you must install separately). You may also register your own classes in the wsgiserver.ssl_adapters dict.

ssl_private_key = None

The filename of the private key to use with SSL.


Start the HTTP server.

statistics = False

Turns statistics-gathering on or off for aware HTTP servers.

thread_pool = 10

The number of worker threads to start up in the pool.

thread_pool_max = -1

The maximum size of the worker-thread pool. Use -1 to indicate no limit.

wsgi_version = (1, 0)

The WSGI version tuple to use with the builtin WSGI server. The provided options are (1, 0) [which includes support for PEP 3333, which declares it covers WSGI version 1.0.1 but still mandates the wsgi.version (1, 0)] and (‘u’, 0), an experimental unicode version. You may create and register your own experimental versions of the WSGI protocol by adding custom classes to the wsgiserver.wsgi_gateways dict.