Native HTTP support

Native HTTP support

HTTPS support (from 1.3)

HTTP sockets

The http-socket <bind> option will make uWSGI natively speak HTTP. If yourweb server does not support the uwsgi protocol but is able tospeak to upstream HTTP proxies, or if you are using a service like Webfactionor Heroku to host your application, you can use http-socket. If you planto expose your app to the world with uWSGI only, use the http optioninstead, as the router/proxy/load-balancer will then be your shield.

The uWSGI HTTP/HTTPS router

uWSGI includes an HTTP/HTTPS router/proxy/load-balancer that can forwardrequests to uWSGI workers. The server can be used in two ways: embedded andstandalone. In embedded mode, it will automatically spawn workers and setupthe communication socket. In standalone mode you have to specify the address ofa uwsgi socket to connect to.

Embedded mode:

./uwsgi --http 127.0.0.1:8080 --master --module mywsgiapp --processes 4

This will spawn a HTTP server on port 8080 that forwards requests to a pool of4 uWSGI workers managed by the master process.

Standalone mode:

./uwsgi --master --http 127.0.0.1:8080 --http-to /tmp/uwsgi.sock

This will spawn a HTTP router (governed by a master for your safety) that willforward requests to the uwsgi socket /tmp/uwsgi.sock. You can bind tomultiple addresses/ports.

[uwsgi]
 
http = 0.0.0.0:8080
http = 192.168.173.17:8181
http = 127.0.0.1:9090
 
master = true
 
http-to = /tmp/uwsgi.sock

And load-balance to multiple nodes:

[uwsgi]
 
http = 0.0.0.0:8080
http = 192.168.173.17:8181
http = 127.0.0.1:9090
 
master = true
 
http-to = /tmp/uwsgi.sock
http-to = 192.168.173.1:3031
http-to = 192.168.173.2:3031
http-to = 192.168.173.3:3031

If you want to go massive (virtualhosting and zero-conf scaling) combine theHTTP router with the uWSGI Subscription Server.
You can make the HTTP server pass custom uwsgi variables to workers with thehttp-var KEY=VALUE option.
You can use the http-modifier1 option to pass a custom modifier1 valueto workers.

HTTPS support

see HTTPS support (from 1.3)

HTTP Keep-Alive

If your backends set the correct HTTP headers, you can use thehttp-keepalive option. Your backends must either set a validContent-Length in each response, or you can use chunked encoding withhttp-auto-chunked. Simply setting “Connection: close” is not enough.

Also remember to set “Connection: Keep-Alive” in your response. You canautomate that using the add-header = Connection: Keep-Alive option.

Since uWSGI 2.1 (master branch) you can use the http11-socket option.http11-socket may replace the add-header and http-keepalive options(but it doesn’t touch tcp stuff as so-keepalive does).Once set the server will try to maintain the connection opened if a bunch ofrules are respected. This is not a smart http 1.1 parser (to avoid parsing thewhole response) but assumes the developer is generating the right headers.http11-socket has been added to support RTSP protocol for video streaming.

HTTP auto gzip

With the http-auto-gzip option, uWSGI can automatically gzip content if theuWSGI-Encoding header is set to gzip while Content-Length andContent-Encoding are not set.

Can I use uWSGI’s HTTP capabilities in production?

If you need a load balancer/proxy it can be a very good idea. It willautomatically find new uWSGI instances and can load balance in various ways.If you want to use it as a real webserver you should take into account thatserving static files in uWSGI instances is possible, but not as good as using adedicated full-featured web server. If you host static assets in the cloud oron a CDN, using uWSGI’s HTTP capabilities you can definitely avoid configuringa full webserver.

Note

If you use Amazon’s ELB (Elastic Load Balancer) in HTTP mode infront of uWSGI in HTTP mode, either a valid Content-Length _must be set_by the backend, or chunked encoding must be used, e.g., withhttp-auto-chunked. The ELB “health test” may still fail in HTTP moderegardless, in which case a TCP health test can be used instead.

Note

In particular, the Django backend does not set Content-Length bydefault, while most others do. If running behind ELB, either use chunkedencoding as above, or force Django to specify Content-Length with theCommonMiddleware (ConditionalGetMiddleware in Django < 1.11)