nginx-代理没有转发到码头-集装箱？

Question

我有一个系统，由两个码头-集装箱，前端和后端。前端容器拥有一个带有颤振web应用程序的nginx-webserver，而后端提供了前端使用的REST接口。前端和后端都是独立的项目，并且都可以独立地工作。两者都在我外部创建的同一个停靠网络(“my”)中。

docker network create my-network

下面是后端的docker-compose.yml：

version: '3'
services:
  my-backend:
    image: my-backend
    ports:
      - "8080:8080"
    networks:
      - my-network
networks:
  my-network:
    driver: bridge
    external: true

这里是最前沿的一个：

version: '3'
services:
  my-frontend:
    image: my-frontend
    ports:
      - "4200:80"
    networks:
      - my-network
networks:
  my-network:
    driver: bridge
    external: true

一切都很好，我可以在我的前端容器中使用curl http://my-backend:8080/somepath。

我的nginx.conf如下：

events {}

http {
    include    /etc/nginx/conf/mime.types;

    server {
        listen 80 default_server;

        root '/usr/share/nginx/html';
        index index.html index.htm;

        location / {
            try_files $uri $uri/ /index.html;
        }

        location /backend {
          proxy_pass http://my-backend:8080;
        }

        location /example {
          proxy_pass https://example.com;
        }
    }
}

正如我所说的，在前端的容器中，我可以使用curl解析my-backend:8080。我还可以从浏览器(使用http://localhost:4200)和示例转发(http://localhost:4200/example)访问前端，但是http://localhost:4200/backend没有到达它的目的地(而http://localhost:8080工作得非常好)。超时后，nginx返回404。有谁能给我解释一下有什么问题吗？

Docker-我使用的图像是最新的nginx和最新的openjdk。

编辑：

我稍微修改了nginx.conf文件，添加

    location /backend {
      resolver 127.0.0.11; # added this line also to http and server-section
      set $backend http://ausanz-backend:8080;
      proxy_pass $backend;
    }

此外，我使用nginx调试。在调试输出中，我看到IP地址被正确解析(调试输出的某些部分)：

my-frontend_1  | 2022/11/08 12:56:02 [debug] 30#30: *5 http request line: "GET /backend/deployment-info HTTP/1.1"
my-frontend_1  | 2022/11/08 12:56:02 [debug] 30#30: *5 http uri: "/backend/deployment-info"
my-frontend_1  | 2022/11/08 12:56:02 [debug] 30#30: *5 test location: "backend"
my-frontend_1  | 2022/11/08 12:56:02 [debug] 30#30: *5 using configuration "/backend"
my-frontend_1  | 2022/11/08 12:56:02 [debug] 30#30: *5 rewrite phase: 3
my-frontend_1  | 2022/11/08 12:56:02 [debug] 30#30: *5 http script value: "http://my-backend:8080"
my-frontend_1  | 2022/11/08 12:56:02 [debug] 30#30: *5 http script set $backend
my-frontend_1  | 2022/11/08 12:56:02 [debug] 30#30: *5 http script var: "http://my-backend:8080"
my-frontend_1  | 2022/11/08 12:56:02 [debug] 30#30: *5 http script copy: "Host"
my-frontend_1  | 2022/11/08 12:56:02 [debug] 30#30: *5 http script var: "my-backend:8080"
my-frontend_1  | 2022/11/08 12:56:02 [debug] 30#30: *5 http script copy: "Connection"
my-frontend_1  | 2022/11/08 12:56:02 [debug] 30#30: *5 http script copy: "close"
my-frontend_1  | "GET /backend/deployment-info HTTP/1.0
my-frontend_1  | Host: my-backend:8080
my-frontend_1  | Connection: close
my-frontend_1  | 2022/11/08 12:56:02 [debug] 30#30: *5 http upstream resolve: "/backend/deployment-info?"
my-frontend_1  | 2022/11/08 12:56:02 [debug] 30#30: *5 name was resolved to 172.19.0.2
my-frontend_1  | 2022/11/08 12:56:02 [debug] 30#30: *5 connect to 172.19.0.2:8080, fd:13 #7
my-frontend_1  | 2022/11/08 12:56:02 [debug] 30#30: *5 http upstream connect: -2
my-frontend_1  | 2022/11/08 12:56:02 [debug] 30#30: *5 http upstream request: "/backend/deployment-info?"
my-frontend_1  | 2022/11/08 12:56:02 [debug] 30#30: *5 http proxy status 404 "404 Not Found"

我证实了curl 172.19.0.2:8080/deployment-info实际上是我的后端。

Answer 1

解决办法相当简单。总是使用path+query参数调用后端。nginx规则没有反映出这一点。解决方案是使用regex作为proxy_pass。不过，解析器仍然需要，否则我得到了502个坏网关。

    location ~ ^/backend/(.*) {
      resolver 127.0.0.11;
      proxy_pass http://my-backend:8080/$1$is_args$args;
    }