Waitress comes bundled with a thin command-line wrapper around the
waitress.serve function called
waitress-serve. This is useful for
development, and in production situations where serving of static assets is
delegated to a reverse proxy, such as Nginx or Apache.
This feature is new as of Waitress 0.8.4.
waitress-serve takes the very same arguments as the
waitress.serve function, but where the function’s arguments have
waitress-serve uses hyphens. Thus:
import myapp waitress.serve(myapp.wsgifunc, port=8041, url_scheme='https')
Is equivalent to:
waitress-serve --port=8041 --url-scheme=https myapp:wsgifunc
The full argument list is given below.
Boolean arguments are represented by flags. If you wish to explicitly set a flag, simply use it by its name. Thus the flag:
Is equivalent to passing
All flags have a negative equivalent. These are prefixed with
using the flag:
Is equivalent to passing
If at any time you want the full argument list, use the
Applications are specified similarly to PasteDeploy, where the format is
myapp.mymodule:wsgifunc. As some application frameworks use application
objects, you can use dots to resolve attributes like so:
A number of frameworks, web.py being an example, have factory methods on
their application objects that return usable WSGI functions when called. For
cases like these,
waitress-serve has the
--call flag. Thus:
waitress-serve --call myapp.mymodule.app.wsgi_factory
Would load the
myapp.mymodule module, and call
app.wsgi_factory to get
a WSGI application function to be passed to
As of 0.8.6, the current directory is automatically included on
waitress-serve [OPTS] MODULE:OBJECT
- Show this information.
- Call the given object to get the WSGI application.
- Hostname or IP address on which to listen, default is ‘0.0.0.0’, which means “all IP addresses on this host”.
- TCP port on which to listen, default is ‘8080’
Path of Unix socket. If a socket path is specified, a Unix domain socket is made instead of the usual inet domain socket.
Not available on Windows.
- Octal permissions to use for the Unix domain socket, default is ‘600’.
wsgi.url_schemevalue, default is ‘http’.
SCRIPT_NAMEWSGI environment value. Setting this to anything except the empty string will cause the WSGI
SCRIPT_NAMEvalue to be the value passed minus any trailing slashes you add, and it will cause the
PATH_INFOof any request which is prefixed with this value to be stripped of the prefix. Default is the empty string.
- Server identity used in the ‘Server’ header in responses. Default is ‘waitress’.
- Number of threads used to process application logic, default is 4.
- Connection backlog for the server. Default is 1024.
- Number of bytes to request when calling
socket.recv(). Default is 8192.
- Number of bytes to send to socket.send(). Default is 18000. Multiples of 9000 should avoid partly-filled TCP packets.
- A temporary file should be created if the pending output is larger than this. Default is 1048576 (1MB).
- A temporary file should be created if the pending input is larger than this. Default is 524288 (512KB).
- Stop creating new channelse if too many are already active. Default is 100.
- Minimum seconds between cleaning up inactive channels. Default is 30. See
- Maximum number of seconds to leave inactive connections open. Default is 120. ‘Inactive’ is defined as ‘has recieved no data from the client and has sent no data to the client’.
- Toggle whether premature client disconnect tracepacks ought to be logged. On by default.
- Maximum size of all request headers combined. Default is 262144 (256KB).
- Maximum size of request body. Default is 1073741824 (1GB).
- Toggle whether to expose tracebacks of unhandled exceptions to the client. Off by default.
- The timeout value in seconds passed to
asyncore.loop(). Default is 1.
- The use_poll argument passed to
asyncore.loop(). Helps overcome open file descriptors limit. Default is False.