How to deploy a WSGI/Python app behind Nginx on Windows?

Question

Coming from the Linux environment, I'll try to be as accurate as possible with Windows concepts I clearly don't master.

Here's my current situation:

I have a functional Windows (10 pro x64) Nginx setup
It's working well on https, using the computer name as FQDN (static files are reachable through https://computer-name/filenames)
My Python Flask app works well and can be served standalone or as WSGI thanks to waitress.

And here's my problem:

I can't figure out how -and if it's even possible- to bind my WSGI app to the server configuration of Nginx. Considering the well written Digital Ocean tutorial for Linux, my problem is probably somewhere between steps 5 and 6. Should I apply the same way of thinking and build a Windows service before linking it thanks to a .sock link, or is that clearly not the right way?

If that's the right way, I don't know where to learn how to do that and what are the main steps to achieve that. For example, it seems that UNIX sockets are now supported by Windows (but I don't know how) or should I use named pipes (but... I don't know how!).

Hopping that I miss nothing, thanks your help.

Edit:

Inspired of this question configuring a proxy pass works. Is that a solid solution?

JFYI nginx on windows should NOT be used for production. “high performance and scalability should not be expected”, “… is considered to be a *beta* version” — Alexey Ten, May 03 '19 at 07:29
May my customer hear you! That's not efficient and the efforts to deploy it are huge. BTW, it's more a proof of concept than a widely deployed app. — Glandalf313, May 03 '19 at 09:20

score 3 · Accepted Answer · answered May 02 '19 at 14:36

According to Waitress Documentation:

unix_socket 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.

Instead of running it from a named pipe you can run it in a local port and reverse proxy that port with nginx. So instead of following DigitalOcean article's step 6 you can do this:

location / {
    proxy_pass http://127.0.0.1:8000;
    proxy_set_header Host $host;
    proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
    proxy_set_header X-Forwarded-Host $host:$server_port;
    proxy_set_header X-Forwarded-Port $server_port;
}

And you can add the waitress serve command to Windows scheduled tasks, see here.

Hope this helps.

Seems to be a really clear & efficient solution. Thanks. The schedule task tip will help too. — Glandalf313, May 02 '19 at 14:40

How to deploy a WSGI/Python app behind Nginx on Windows?

Edit:

1 Answers1