NGINX target specific upload files on different folders

Question

I need to add a noindex tag to multiple specific uploaded files, but they all belong to different year and month upload folders like this for example:

• /var/www/public/app/uploads/2021/06/file-1.pdf
• /var/www/public/app/uploads/2018/11/file-2.pdf
• /var/www/public/app/uploads/2011/07/file-3.pdf

I did this initially, but did not work:

location ~ "^(.*)/app/uploads/(.*)$" {
    location ~ ^/(file-1.pdf|file-2.pdf|file-3.pdf)$ {
        add_header X-Robots-Tag "noindex";
    }
}

I don't have access yet to test, but this is what I'm planning to do instead:

location ~ "/var/www/public/app/uploads/(.*)$" {
    location ~ ^/(file-1.pdf|file-2.pdf|file-3.pdf)$ {
        add_header X-Robots-Tag "noindex";
    }
}

I wonder if there is a better approach, or if this will even work. Right now there are 27 specific files I have to do this to so I'm not sure if the file-1.pdf|file-2.pdf|file-3.pdf is my best option.

Any help is appreciated, thanks.

Answer 1

First of all, you are definitely use the location directive incorrectly. Assuming your server root directory is /var/www/public/app and your sample request is http://example.com/uploads/2021/06/file-1.pdf the normalized URI (which is subject to check for location or rewrite directives) will be /uploads/2021/06/file-1.pdf. An example of location that will catch those requests and add a required header is

location ~ ^/uploads/(?<file>.*) {
    if ($file ~ ^(?:2021/06/file-1\.pdf|2018/11/file-2\.pdf|2011/07/file-3\.pdf))$ {
        add_header X-Robots-Tag "noindex";
    }
}

Here (?<file>.*) is so-called named capture group for later usage. Using two nested locations is possible too:

location /uploads/ {
    location ~ /(?:2021/06/file-1\.pdf|2018/11/file-2\.pdf|2011/07/file-3\.pdf)$ {
        add_header X-Robots-Tag "noindex";
    }
}

You can also use map directive (although I don't know if this could be called "a better approach"):

map $file $robots {
    2021/06/file-1.pdf    noindex;
    2018/11/file-2.pdf    noindex;
    2011/07/file-3.pdf    noindex;
    # default value will be an empty string
}

server {
    ...
    location ~ ^/uploads/(?<file>.*) {
        add_header X-Robots-Tag $robots;
    }
    ...
}

or without regexes:

map $uri $robots {
    /uploads/2021/06/file-1.pdf    noindex;
    /uploads/2018/11/file-2.pdf    noindex;
    /uploads/2011/07/file-3.pdf    noindex;
    # default value will be an empty string
}

server {
    ...
    location /uploads/ {
        add_header X-Robots-Tag $robots;
    }
    ...
}

The add_header directive won't add a header to response if its second parameter will be an empty string.

You should also take in attention this documentation excerpt:

These directives are inherited from the previous configuration level if and only if there are no add_header directives defined on the current level.