Accepting the PROXY Protocol
This article explains how to configure NGINX and F5 NGINX Plus to accept the PROXY protocol, rewrite the IP address of a load balancer or proxy to the one received in the PROXY protocol header, configure simple logging of a client’s IP address, and enable the PROXY protocol between NGINX and a TCP upstream server.
Introduction
The PROXY protocol enables NGINX and NGINX Plus to receive client connection information passed through proxy servers and load balancers such as HAproxy and Amazon Elastic Load Balancer (ELB).
With the PROXY protocol, NGINX can learn the originating IP address from HTTP, SSL, HTTP/2, SPDY, WebSocket, and TCP. Knowing the originating IP address of a client may be useful for setting a particular language for a website, keeping a denylist of IP addresses, or simply for logging and statistics purposes.
The information passed via the PROXY protocol is the client IP address, the proxy server IP address, and both port numbers.
Using this data, NGINX can get the originating IP address of the client in several ways:
-
With the
$proxy_protocol_addr
and$proxy_protocol_port
variables which capture the original client IP address and port. The$remote_addr
and$remote_port
variables capture the IP address and port of the load balancer. -
With the RealIP module which rewrites the values in the
$remote_addr
and$remote_port
variables, replacing the IP address and port of the load balancer with the original client IP address and port. The$realip_remote_addr
and$realip_remote_port
variables retain the address and port of the load balancer, and the$proxy_protocol_addr
and$proxy_protocol_port
variables retain the original client IP address and port anyway.
Prerequisites
-
To accept the PROXY protocol v2, NGINX Plus R16 and later or NGINX Open Source 1.13.11 and later
-
To accept the PROXY protocol for HTTP, NGINX Plus R3 and later or NGINX Open Source 1.5.12 and later
-
For TCP client‑side PROXY protocol support, NGINX Plus R7 and later or NGINX Open Source 1.9.3 and later
-
To accept the PROXY protocol for TCP, NGINX Plus R11 and later or NGINX Open Source 1.11.4 and later
-
The Real‑IP modules for HTTP and Stream TCP are not included in NGINX Open Source by default; see Installing NGINX Open Source for details. No extra steps are required for NGINX Plus.
Configuring NGINX to Accept the PROXY Protocol
To configure NGINX to accept PROXY protocol headers, add the proxy_protocol
parameter to the listen
directive in a server
block in the http {}
or stream {}
block.
http {
#...
server {
listen 80 proxy_protocol;
listen 443 ssl proxy_protocol;
#...
}
}
stream {
#...
server {
listen 12345 proxy_protocol;
#...
}
}
Now you can use the $proxy_protocol_addr
and $proxy_protocol_port
variables for the client IP address and port and additionally configure the HTTP and stream
RealIP modules to replace the IP address of the load balancer in the $remote_addr
and $remote_port
variables with the IP address and port of the client.
Changing the Load Balancer’s IP Address To the Client IP Address
You can replace the address of the load balancer or TCP proxy with the client IP address received from the PROXY protocol. This can be done with the HTTP and stream
RealIP modules. With these modules, the $remote_addr
and $remote_port
variables retain the real IP address and port of the client, while the $realip_remote_addr
and $realip_remote_port
variables retain the IP address and port of the load balancer.
To change the IP address from the load balancer’s IP address to the client’s IP address:
-
Make sure you’ve configured NGINX to accept the PROXY protocol headers. See Configuring NGINX to Accept the PROXY Protocol.
-
Make sure that your NGINX installation includes the HTTP and Stream Real‑IP modules:
nginx -V 2>&1 | grep -- 'http_realip_module' nginx -V 2>&1 | grep -- 'stream_realip_module'
If not, recompile NGINX with these modules. See Installing NGINX Open Source for details. No extra steps are required for NGINX Plus.
-
In the
set_real_ip_from
directive for HTTP, Stream, or both, specify the IP address or the CIDR range of addresses of the TCP proxy or load balancer:server { #... set_real_ip_from 192.168.1.0/24; #... }
-
In the
http {}
context, change the IP address of the load balancer to the IP address of the client received from the PROXY protocol header, by specifying theproxy_protocol
parameter to thereal_ip_header
directive:http { server { #... real_ip_header proxy_protocol; } }
Logging the Original IP Address
When you know the original IP address of the client, you can configure the correct logging:
-
For HTTP, configure NGINX to pass the client IP address to upstream servers using the
$proxy_protocol_addr
variable with theproxy_set_header
directive:http { proxy_set_header X-Real-IP $proxy_protocol_addr; proxy_set_header X-Forwarded-For $proxy_protocol_addr; }
-
Add the
$proxy_protocol_addr
variable to thelog_format
directive (HTTP or Stream):-
In the
http
block:http { #... log_format combined '$proxy_protocol_addr - $remote_user [$time_local] ' '"$request" $status $body_bytes_sent ' '"$http_referer" "$http_user_agent"'; }
-
In the
stream
block:stream { #... log_format basic '$proxy_protocol_addr - $remote_user [$time_local] ' '$protocol $status $bytes_sent $bytes_received ' '$session_time'; }
-
PROXY Protocol for a TCP Connection to an Upstream
For a TCP stream, the PROXY protocol can be enabled for connections between NGINX and an upstream server. To enable the PROXY protocol, include the proxy_protocol
directive in a server
block at the stream {}
level:
stream {
server {
listen 12345;
proxy_pass example.com:12345;
proxy_protocol on;
}
}
Example
http {
log_format combined '$proxy_protocol_addr - $remote_user [$time_local] '
'"$request" $status $body_bytes_sent '
'"$http_referer" "$http_user_agent"';
#...
server {
server_name localhost;
listen 80 proxy_protocol;
listen 443 ssl proxy_protocol;
ssl_certificate /etc/nginx/ssl/public.example.com.pem;
ssl_certificate_key /etc/nginx/ssl/public.example.com.key;
location /app/ {
proxy_pass http://backend1;
proxy_set_header Host $host;
proxy_set_header X-Real-IP $proxy_protocol_addr;
proxy_set_header X-Forwarded-For $proxy_protocol_addr;
}
}
}
stream {
log_format basic '$proxy_protocol_addr - $remote_user [$time_local] '
'$protocol $status $bytes_sent $bytes_received '
'$session_time';
#...
server {
listen 12345 ssl proxy_protocol;
ssl_certificate /etc/nginx/ssl/cert.pem;
ssl_certificate_key /etc/nginx/ssl/cert.key;
proxy_pass backend.example.com:12345;
proxy_protocol on;
}
}
The example assumes that there is a load balancer in front of NGINX to handle all incoming HTTPS traffic, for example Amazon ELB. NGINX accepts HTTPS traffic on port 443 (listen 443 ssl;
), TCP traffic on port 12345, and accepts the client’s IP address passed from the load balancer via the PROXY protocol as well (the proxy_protocol
parameter to the listen
directive in both the http {}
and stream {}
blocks.
NGINX terminates HTTPS traffic (the ssl_certificate
and ssl_certificate_key
directives) and proxies the decrypted data to a backend server:
- For HTTP:
proxy_pass http://backend1;
- For TCP:
proxy_pass backend.example.com:12345
It includes the client IP address and port with the proxy_set_header
directives.
The $proxy_protocol_addr
variable specified in the log_format
directive also passes the client’s IP address to the log for both HTTP and TCP.
Additionally, a TCP server (the stream {}
block) sends its own PROXY protocol data to its backend servers (the proxy_protocol on
directive).