Skip to content

Identifying an original client IP address if using an HTTP proxy or a load balancer (NGINX)

These instructions describe the NGINX configuration required to identify an originating IP address of a client connecting to your servers through an HTTP proxy or load balancer. This is relevant for the self-hosted NGINX-based nodes.

  • If the self-hosted Wallarm node is installed from the all-in-one installer, AWS / GCP images or the NGINX-based Docker image, please use the current instructions.

  • If the self-hosted Wallarm node is deployed as the K8s Ingress controller, please use these instructions.

How Wallarm node identifies an IP address of a request

The Wallarm node reads a request source IP address from the NGINX variable $remote_addr. If the request passed through a proxy server or load balancer before being sent to the node, the variable $remote_addr retains the proxy server or load balancer IP address.

Using balancer

The request source IP address identified by the Wallarm node is displayed in the attack details in Wallarm Console.

Possible problems of using a proxy server or load balancer IP address as a request source address

If the Wallarm node considers the proxy server or load balancer IP address to be the IP address of the request source, the following Wallarm features may work incorrectly:

  • Controlling access to applications by IP addresses, for example:

    If original client IP addresses were denylisted, the Wallarm node still would not block requests originated from them since it considers the load balancer IP address to be the IP address of the request source.

  • Brute force protection, for example:

    If requests passed through the load balancer have brute force attack signs, Wallarm will denylist this load balancer IP address and therefore will block all further requests passed through this load balancer.

  • The Active threat verification module and Vulnerability Scanner, for example:

    Wallarm will consider the load balancer IP address to be the IP address originating test attacks generated by the Active threat verification module and Vulnerability Scanner. Thus, test attacks will be displayed in Wallarm Console as attacks originated from the load balancer IP address and will be additionally checked by Wallarm which will create an extra load on the application.

If the Wallarm node is connected via an IPC socket, then 0.0.0.0 will be considered as a request source.

Configuration for an original client IP address identification

To configure an original client IP address identification, you can use the NGINX module ngx_http_realip_module. This module allows redefining the value of $remote_addr used by the Wallarm node to get a client IP address.

You can use the NGINX module ngx_http_realip_module in one of the following ways:

  • To read an original client IP address from a specific header (usually, X-Forwarded-For) added to the request by a load balancer or proxy server.

  • If a load balancer or proxy server supports the PROXY protocol, to read an original client IP address from the header PROXY.

Configuring NGINX to read the header X-Forwarded-For (X-Real-IP or a similar)

If a load balancer or proxy server appends a header X-Forwarded-For (X-Real-IP or a similar) containing an original client IP address, please configure the NGINX module ngx_http_realip_module to read this header as follows:

  1. Open the following configuration file of NGINX installed with the Wallarm node:

    • /etc/nginx/sites-enabled/default if the Wallarm node is installed from the all-in-one installer or the AWS / GCP image.
    • If the Wallarm node is deployed from the NGINX-based Docker image, you should create and edit the NGINX configuration file locally and mount it to the Docker container at the path /etc/nginx/sites-enabled/default. You can copy an initial NGINX configuration file and get the instructions on mounting the file to the container from the instructions on the Wallarm NGINX-based Docker instructions.

  2. In the NGINX context location or higher, add the directive set_real_ip_from with a proxy server or load balancer IP address. If a proxy server or load balancer has several IP addresses, please add an appropriate number of separate directives. For example:

    ...
location / {
    wallarm_mode block;

    set_real_ip_from 1.2.3.4;
    set_real_ip_from 192.0.2.0/24;
}
...

  3. In the documentation on a load balancer being used, find the name of the header appended by this load balancer to pass an original client IP address. Most frequently, the header is called X-Forwarded-For.

  4. In the NGINX context location or higher, add the directive real_ip_header with the header name found in the previous step. For example:

    ...
location / {
    wallarm_mode block;

    set_real_ip_from 1.2.3.4;
    set_real_ip_from 192.0.2.0/24;
    real_ip_header X-Forwarded-For;
}
...

  5. Restart NGINX:

    sudo systemctl restart nginx

    sudo service nginx restart

    sudo systemctl restart nginx

    sudo systemctl restart nginx

    sudo systemctl restart nginx

    NGINX will assign the value of the header specified in the real_ip_header directive to the $remote_addr variable, so the Wallarm node will read original client IP addresses from this variable.

  6. Test the configuration.

Configuring NGINX to read the header PROXY

If a load balancer or proxy server supports the PROXY protocol, you can configure the NGINX module ngx_http_realip_module to read the header PROXY as follows:

  1. Open the following configuration file of NGINX installed with the Wallarm node:

    • /etc/nginx/sites-enabled/default if the Wallarm node is installed from the all-in-one installer or the AWS / GCP image.
    • If the Wallarm node is deployed from the NGINX-based Docker image, you should create and edit the NGINX configuration file locally and mount it to the Docker container at the path /etc/nginx/sites-enabled/default. You can copy an initial NGINX configuration file and get the instructions on mounting the file to the container from the instructions on the Wallarm NGINX-based Docker instructions.

  2. In the NGINX context server, add the parameter proxy_protocol to the directive listen.

  3. In the NGINX context location or higher, add the directive set_real_ip_from with a proxy server or load balancer IP address. If a proxy server or load balancer has several IP addresses, please add an appropriate number of separate directives. For example:

  4. In the NGINX context location or higher, add the directive real_ip_header with the value of proxy_protocol.

    An example NGINX configuration file with all directives added:

    server {
    listen 80 proxy_protocol;
    server_name localhost;

    set_real_ip_from <IP_ADDRESS_OF_YOUR_PROXY>;
    real_ip_header proxy_protocol;

    ...
}
    • NGINX listens for incoming connections on port 80.
    • If the header PROXY is not passed in the incoming request, NGINX does not accept this request as it is considered not valid.
    • For requests originated from the address <IP_ADDRESS_OF_YOUR_PROXY>, NGINX assigns the source address passed in the header PROXY to the variable $remote_addr, so the Wallarm node will read original client IP addresses from this variable.

  5. Restart NGINX:

    sudo systemctl restart nginx

    sudo service nginx restart

    sudo systemctl restart nginx

    sudo systemctl restart nginx

    sudo systemctl restart nginx

  6. Test the configuration.

To include an original client IP address in the logs, you should add the directive proxy_set_header and edit the list of variables in the log_format directive in the NGINX configuration as described in the NGINX logging instructions.

More details on identifying an original client IP address based on the PROXY header are available in the NGINX documentation.

Testing the configuration

  1. Send the test attack to the protected application address:

    curl http://localhost/etc/passwd

    printf "PROXY TCP4 <IP_ADDRESS_OF_YOUR_PROXY> <REAL_CLIENT_IP> 0 80\r\nGET /etc/passwd\r\n\r\n" | nc localhost 80

  2. Open Wallarm Console and make sure that the original client IP address is displayed in the attack details:

    IP address originated the request

    If NGINX read the original address from the header X-Forwarded-For (X-Real-IP or a similar), the header value would also be displayed in the raw attack.

    Header X-Forwarded-For

Configuration examples

Below you will find examples of the NGINX configuration required to identify an originating IP address of a client connecting to your servers through popular load balancers.

Cloudflare CDN

If using Cloudflare CDN, you can configure the NGINX module ngx_http_realip_module to identify original client IP addresses.

...
set_real_ip_from 103.21.244.0/22;
set_real_ip_from 103.22.200.0/22;
set_real_ip_from 103.31.4.0/22;
set_real_ip_from 104.16.0.0/12;
set_real_ip_from 108.162.192.0/18;
set_real_ip_from 131.0.72.0/22;
set_real_ip_from 141.101.64.0/18;
set_real_ip_from 162.158.0.0/15;
set_real_ip_from 172.64.0.0/13;
set_real_ip_from 173.245.48.0/20;
set_real_ip_from 188.114.96.0/20;
set_real_ip_from 190.93.240.0/20;
set_real_ip_from 197.234.240.0/22;
set_real_ip_from 198.41.128.0/17;
set_real_ip_from 2400:cb00::/32;
set_real_ip_from 2606:4700::/32;
set_real_ip_from 2803:f800::/32;
set_real_ip_from 2405:b500::/32;
set_real_ip_from 2405:8100::/32;
set_real_ip_from 2c0f:f248::/32;
set_real_ip_from 2a06:98c0::/29;

real_ip_header CF-Connecting-IP;
#real_ip_header X-Forwarded-For;
real_ip_recursive on;
...

  • Before saving the configuration, please ensure that Cloudflare IP addresses specified in the configuration above match those in the Cloudflare documentation.

  • In the real_ip_header directive value, you can specify either CF-Connecting-IP or X-Forwarded-For. Cloudflare CDN appends both headers and you can configure NGINX to read any of them. More details in the Cloudflare CDN

Fastly CDN

If using Fastly CDN, you can configure the NGINX module ngx_http_realip_module to identify original client IP addresses.

...
set_real_ip_from 23.235.32.0/20;
set_real_ip_from 43.249.72.0/22;
set_real_ip_from 103.244.50.0/24;
set_real_ip_from 103.245.222.0/23;
set_real_ip_from 103.245.224.0/24;
set_real_ip_from 104.156.80.0/20;
set_real_ip_from 146.75.0.0/16;
set_real_ip_from 151.101.0.0/16;
set_real_ip_from 157.52.64.0/18;
set_real_ip_from 167.82.0.0/17;
set_real_ip_from 167.82.128.0/20;
set_real_ip_from 167.82.160.0/20;
set_real_ip_from 167.82.224.0/20;
set_real_ip_from 172.111.64.0/18;
set_real_ip_from 185.31.16.0/22;
set_real_ip_from 199.27.72.0/21;
set_real_ip_from 199.232.0.0/16;
set_real_ip_from 2a04:4e40::/32;
set_real_ip_from 2a04:4e42::/32;

real_ip_header X-Forwarded-For;
real_ip_recursive on;
...

Before saving the configuration, please ensure that Fastly IP addresses specified in the configuration above match those in the Fastly documentation.

HAProxy

If using HAProxy, both HAProxy and Wallarm node sides should be properly configured to identify original client IP addresses:

  • In the /etc/haproxy/haproxy.cfg configuration file, insert the option forwardfor header X-Client-IP line into the backend directive block responsible for connecting HAProxy to the Wallarm node.

    The option forwardfor directive tells the HAProxy balancer that a header with the IP address of the client must be added to the request. More details in the HAProxy documentation

    Configuration example:

    ...
# Public IP address for receiving requests
frontend my_frontend
    bind <HAPROXY_IP>
    mode http
    default_backend my_backend

# Backend with the Wallarm node
backend my_backend
    mode http
option forwardfor header X-Client-IP
server wallarm-node <WALLARM_NODE_IP>
...
    • <HAPROXY_IP> is the IP address of the HAProxy server to receive client requests.
    • <WALLARM_NODE_IP> is the IP address of the Wallarm node to receive requests from the HAProxy server.

  • In the configuration file of NGINX installed with the Wallarm node, configure the module ngx_http_realip_module as follows:

    ...
location / {
    wallarm_mode block;

    proxy_pass http://<APPLICATION_IP>;        
    set_real_ip_from <HAPROXY_IP1>;
    set_real_ip_from <HAPROXY_IP2>;
    real_ip_header X-Client-IP;
}
...
    • <APPLICATION_IP> is the IP address of the protected application for requests from the Wallarm node.
    • <HAPROXY_IP1> and <HAPROXY_IP2> are IP addresses of HAProxy balancers that pass requests to the Wallarm node.