Nginx代理后如何针对请求的Client IP进行限流

## 因为使用了Nginx，针对请求的Client IP进行限流时有问题

`所有请求获取到的ip地址都为同一个地址172.*.*.*，（应用部署在docker容器内）。`

看了下com.alibaba.csp.sentinel.adapter.gateway.sc.ServerWebExchangeItemParser#getRemoteAddress的方法，获取ip时使用的代码是
```
public String getRemoteAddress(ServerWebExchange exchange) {
InetSocketAddress remoteAddress = exchange.getRequest().getRemoteAddress();
return remoteAddress == null ? null : remoteAddress.getAddress().getHostAddress();
}
```
是否可以自定义获取请求ip地址？或者可以让sentinel得到代理后真实请求ip的方法

Start with com.alibaba.csp.sentinel.adapter.gateway.sc.ServerWebExchangeItemParser#getRemoteAddress and reproduce the reported client-IP behavior through Nginx into the Docker-deployed application. Check how the gateway adapter obtains the request address and determine whether proxy-aware or configurable address handling is supported; done means requests are rate-limited by the real client IP rather than the shared 172.*.*.* address.

Written by the indexing model from the issue text.