> On 6 May 2022, at 03:54, Alan Jackson <alan.jackson@netvalue.nz> wrote:
>
> Hi,
>
> I'm trying to use rate limiting on an nginx mail proxy->nginx fastcgi backend to restrict the number of concurrent connections from a client's IP. Unfortunately, I can't use proxy_protocol on the mail proxy side due to the nginx version not supporting it.
>
> Using debug logging, I know that the nginx mail proxy is passing through an http header 'Client-IP' which contains the originating client IP.
>
> I've confirmed that in the http backend, the variable $http_client_ip contains the originating request's client ip ($binary_remote_addr just contains 127.0.0.1 due to the proxy), and that setting a return http header using this variable works just fine.
>
> However, when I try to use limit_conn_zone with the $http_client_ip variable, it is not behaving as I expect:
>
> "limit_conn_zone $http_client_ip zone=perip:10m;"
>
> Results in log lines like:
>
> "limiting connections, dry run, by zone "perip", client: 127.0.0.1, server: _, request: "GET /mail_auth.php HTTP/1.0"
>
> which doesn't look like the variable is being instantiated properly (client is still 127.0.0.1).
>

The "client: 127.0.0.1" part is the above log represents client address
you previously claimed to contain "127.0.0.1", which is fine.
The limit_conn_zone key evaluated content is logged at the debug level.

> Is using an $http_ variable expected to work in a limit_conn_zone config like this?
>

Yes, it is.

_______________________________________________
nginx mailing list -- nginx@nginx.org
To unsubscribe send an email to nginx-leave@nginx.org