Arguments to waitress.serveΒΆ

Here are the arguments you can pass to the waitress.serve` function or use in PasteDeploy configuration (interchangeably):


Hostname or IP address (string) on which to listen, default, which means "all IP addresses on this host".


May not be used with listen


TCP port (integer) on which to listen, default 8080


May not be used with listen


Tell waitress to listen on combinations of host:port arguments. Combinations should be a quoted, space-delimited list, as in the following examples.

listen=" [::1]:8080"
listen="*:8080 *:6543"

A wildcard for the hostname is also supported and will bind to both IPv4/IPv6 depending on whether they are enabled or disabled.

IPv6 IP addresses are supported by surrounding the IP address with brackets.

New in version 1.0.

Enable or disable IPv4 (boolean)
Enable or disable IPv6 (boolean)

Path of Unix socket (string). If a socket path is specified, a Unix domain socket is made instead of the usual inet domain socket.

Not available on Windows.

Default: None


Octal permissions to use for the Unix domain socket (string). Only used if unix_socket is not None.

Default: '600'


A list of sockets. The sockets can be either Internet or UNIX sockets and have to be bound. Internet and UNIX sockets cannot be mixed. If the socket list is not empty, waitress creates one server for each socket.

Default: []

New in version 1.1.1.


May not be used with listen, host, port or unix_socket


The number of threads used to process application logic (integer).

Default: 4


IP address of a remote peer allowed to override various WSGI environment variables using proxy headers.

For unix sockets, set this value to localhost instead of an IP address.

Default: None


How many proxies we trust when chained. For example,

X-Forwarded-For:, "[2001:db8::1]"


Forwarded: for=, For="[2001:db8::1]"

means there were (potentially), two proxies involved. If we know there is only 1 valid proxy, then that initial IP address "" is not trusted and we completely ignore it.

If there are two trusted proxies in the path, this value should be set to 2. If there are more proxies, this value should be set higher.

Default: 1

New in version 1.2.0.


Which of the proxy headers should we trust, this is a set where you either specify "forwarded" or one or more of "x-forwarded-host", "x-forwarded-for", "x-forwarded-proto", "x-forwarded-port", "x-forwarded-by".

This list of trusted headers is used when trusted_proxy is set and will allow waitress to modify the WSGI environment using the values provided by the proxy.

New in version 1.2.0.


If trusted_proxy is set, the default is x-forwarded-proto to match older versions of Waitress. Users should explicitly opt-in by selecting the headers to be trusted as future versions of waitress will use an empty default.


It is an error to set this value without setting trusted_proxy.


Should waitress log warning messages about proxy headers that are being sent from upstream that are not trusted by trusted_proxy_headers but are being cleared due to clear_untrusted_proxy_headers?

This may be useful for debugging if you expect your upstream proxy server to only send specific headers.

Default: False

New in version 1.2.0.


It is a no-op to set this value without also setting clear_untrusted_proxy_headers and trusted_proxy


This tells Waitress to remove any untrusted proxy headers ("Forwarded", "X-Forwared-For", "X-Forwarded-By", "X-Forwarded-Host", "X-Forwarded-Port", "X-Forwarded-Proto") not explicitly allowed by trusted_proxy_headers.

Default: False

New in version 1.2.0.


The default value is set to False for backwards compatibility. In future versions of Waitress this default will be changed to True. Warnings will be raised unless the user explicitly provides a value for this option, allowing the user to opt-in to the new safety features automatically.


It is an error to set this value without setting trusted_proxy.


The value of wsgi.url_scheme in the environ. This can be overridden per-request by the value of the X_FORWARDED_PROTO header, but only if the client address matches trusted_proxy.

Default: http


Server identity (string) used in "Server:" header in responses.

Default: waitress


The value waitress passes to pass to socket.listen() (integer). This is the maximum number of incoming TCP connections that will wait in an OS queue for an available channel. From listen(1): "If a connection request arrives when the queue is full, the client may receive an error with an indication of ECONNREFUSED or, if the underlying protocol supports retransmission, the request may be ignored so that a later reattempt at connection succeeds."

Default: 1024


The argument waitress passes to socket.recv() (integer).

Default: 8192


The number of bytes to send to socket.send() (integer). Multiples of 9000 should avoid partly-filled TCP packets, but don't set this larger than the TCP write buffer size. In Linux, /proc/sys/net/ipv4/tcp_wmem controls the minimum, default, and maximum sizes of TCP write buffers.

Default: 18000


A tempfile should be created if the pending output is larger than outbuf_overflow, which is measured in bytes. The default is conservative.

Default: 1048576 (1MB)


A tempfile should be created if the pending input is larger than inbuf_overflow, which is measured in bytes. The default is conservative.

Default: 524288 (512K)


Stop creating new channels if too many are already active (integer). Each channel consumes at least one file descriptor, and, depending on the input and output body sizes, potentially up to three, plus whatever file descriptors your application logic happens to open. The default is conservative, but you may need to increase the number of file descriptors available to the Waitress process on most platforms in order to safely change it (see ulimit -a "open files" setting). Note that this doesn't control the maximum number of TCP connections that can be waiting for processing; the backlog argument controls that.

Default: 100


Minimum seconds between cleaning up inactive channels (integer). See also channel_timeout.

Default: 30


Maximum seconds to leave an inactive connection open (integer). "Inactive" is defined as "has received no data from a client and has sent no data to a client".

Default: 120


Set to False to not log premature client disconnect tracebacks.

Default: True


Maximum number of bytes of all request headers combined (integer).

Default: 262144 (256K)


Maximum number of bytes in request body (integer).

Default: 1073741824 (1GB)


Set to True to expose tracebacks of unhandled exceptions to client.

Default: False


The timeout value (seconds) passed to asyncore.loop to run the mainloop.

Default: 1

New in version 0.8.3.


Set to True to switch from using select() to poll() in asyncore.loop. By default asyncore.loop() uses select() which has a limit of 1024 file descriptors. select() and poll() provide basically the same functionality, but poll() doesn't have the file descriptors limit.

Default: False

New in version 0.8.6.


String: the value used as the WSGI SCRIPT_NAME value. Setting this to anything except the empty string will cause the WSGI SCRIPT_NAME value to be the value passed minus any trailing slashes you add, and it will cause the PATH_INFO of any request which is prefixed with this value to be stripped of the prefix.

Default: ''