$http_origin in Nginx config files

$http_<headername> is automatically created based on the request headers. You can see the same on

https://nginx.org/en/docs/http/ngx_http_core_module.html#var_http_

$http_name

arbitrary request header field; the last part of a variable name is the field name converted to lower case with dashes replaced by underscores

If you look closer at http://nginx.org/docs/varindex.html, you'll notice there's only a single item with a prefix of $http_ mentioned on that page; and, in fact, this single item is lacking any further suffix, being generic on purpose.

The functionality is documented over at http://nginx.org/r/$http_ as follows:

$http_name — arbitrary request header field; the last part of a variable name is the field name converted to lower case with dashes replaced by underscores

---

However, if you look into the source code behind nginx, it does have internal optimisations for storing certain specific request headers in individual variables for performance reasons (these are defined as a list in the code below, but they are henceforth placed into a hash table during the initialisation process of nginx by ngx_http_variables_add_core_vars() function). For example, $http_host, $http_user_agent and $http_referer, plus a bunch of other variables that may have to be handled internally by certain nginx components, or which may be frequently used within the configuration files.

• http://ngx.su/src/http/ngx_http_request.h#ngx_http_header_t
• http://ngx.su/src/http/ngx_http_variables.c#ngx_http_core_variables
• http://ngx.su/src/http/ngx_http_variables.c#ngx_http_variables_add_core_vars

However, as an end-user, you don't really have to worry about those kinds of details — the generic $http_name is sufficient to describe the whole functionality. (In fact, $http_origin isn't special or popular enough to have warranted its own specialised handling.)