Redirect to external asset nginx

Question

My goal is to have one nginx that can proxy_pass to other server.

Desired input https://example.com/https://assets1.com/image.jpg?utm=whatever Desired output https://assets1.com/image.jpg?utm=whatever 

Here my location block

server { index index.html index.htm index.nginx-debian.html; server_name example.com www.example.com; location ~/(.*) { if ($request_method = 'GET') { add_header 'Access-Control-Allow-Origin' '*'; add_header 'Access-Control-Allow-Credentials' 'true'; add_header 'Access-Control-Allow-Methods' 'GET, POST, OPTIONS'; add_header 'Access-Control-Allow-Headers' 'DNT,X-CustomHeader,Keep-Alive,User-> } proxy_pass https://$1$is_args$args/; proxy_set_header X-Real-IP $remote_addr; proxy_set_header Host "www.example.com"; } listen 80; listen [::]:442 ssl ipv6only=on; listen 443 ssl; # managed by Certbot ssl_certificate /etc/letsencrypt/live/example.com/fullchain.pem; # managed by Certbot ssl_certificate_key /etc/letsencrypt/live/example.com/privkey.pem; # managed by Certbot include /etc/letsencrypt/options-ssl-nginx.conf; # managed by Certbot ssl_dhparam /etc/letsencrypt/ssl-dhparams.pem; # managed by Certbot } 

And the error I get:

2021/09/21 09:27:32 [error] 8475#8475: *16 invalid port in upstream "https:/assets1.com/image.jpg?utm=whatever", client: [IP], server: domain.com, request: "GET /https://assets1.com/image.jpg?utm=whatever HTTP/1.1", host: "example.com"

Answer 1

Your original request URL contains protocol prefix: https://example.com/https://assets1.com.

Your location block captures the part after first /, so $1 becomes https://assets1.com.

In your proxy_pass statement you have https://$1$is_args$args, which becomes https://https://assets1.com when variable is expanded.

nginx tries to parse https://assets1.com as an domain:port pair, so the domain part of URL is https and port is an empty string.

To address the issue, I propose the following configuration:

location ~^/https://(.+)$ { proxy_pass https://$1$is_args$args; ... } 

This way we exclude the protocol part from being captured into $1, so that we have a proper URL. I also added start and end anchors to make the regex more robust.