realip | golang library that can get client | Networking library

For your case, you will need to add a unique index on ip (if it doesn't exist) column and then you can do like this:

The application implicitly sets the cookie path to the login handler path. Fix by explicitly setting the cookie path to "/".

If that rewrite rule had a break flag instead of last, it would remove the /STFlow prefix from /STFlow/some/path URI before passing it to the upstream, i. e. do the same as the second location block except of setting debug-header and $realip variable. But as far as I understand using the last flag makes that four lines never executed, further URI processing would be done inside the second location block.

As far as I understood nginx documentation only one location should be selected and executed

Only a single location is indeed selected for serving a request. It does not mean that there is no "jumping" done between the locations (contexts) while reaching the final location.

This is what the rewrite module is actually responsible for: jumping between contexts by rewriting current URI / repeating NGINX's location search based on the rewritten URI.

NGINX first finds the location with longest prefix for serving request. Specific to your example, this is location /STFlow/ {. Now it selected it for evaluation.

It sees rewrite ^/STFlow(.*)$ $1 last; which modifies current URI to /dashboard. The last keyword to rewrite directive triggers location search all over again based on the now current /dashboard URI. Worth noting that while jumping location /STFlow/ { to location / { the directives from the former are not going to apply. After the "jump", now location / { is the one currently selected for serving the request, etc.

The location / { has no more directives (no try_files, rewrite, etc.) that can modify the current URI and trigger searching all over again. Thus it is the final location that will be used by NGINX for constructing the response.

The directives/configuration that will finally apply to serving response is the ones from the final location or its parent, if the final location is a nested location. Even then, it depends on specific directives. Some directives will be inherited from parent location and some would not. It is a major topic of its own. For example:

Unless you want to write your own parser, you can just easily add the JSON package from org.json. Either download it or just add it as a dependency. It's fairly straight forward to use and there is a good reference here: https://www.baeldung.com/java-org-json

as for recovering the json from a (I'm assuming external) site you could write something like this: