如何配置Kubernetes中Nginx前端以实现多路径路由转发

Got it, let's walk through exactly how to set up an Nginx reverse proxy in Kubernetes that routes requests to your app1 and app2 services via /app1 and /app2 paths. I'll break this down into actionable steps with concrete examples.

First, you'll need a custom Nginx config that maps each path to the corresponding backend service. Create a file named nginx.conf with this content:

events {}

http {
    server {
        listen 80;
        server_name localhost;

        # Route /app1 requests to app1 service
        location /app1/ {
            # Rewrite the path to strip /app1/ so the backend gets the root path
            rewrite ^/app1/(.*)$ /$1 break;
            proxy_pass http://app1:80/;
            proxy_set_header Host $host;
            proxy_set_header X-Real-IP $remote_addr;
            proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
        }

        # Route /app2 requests to app2 service
        location /app2/ {
            rewrite ^/app2/(.*)$ /$1 break;
            proxy_pass http://app2:80/;
            proxy_set_header Host $host;
            proxy_set_header X-Real-IP $remote_addr;
            proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
        }

        # Optional: Default route if someone hits the root
        location / {
            return 200 "Welcome to the Nginx Proxy!\nUse /app1 or /app2 to access services.\n";
        }
    }
}

What's happening here? #

• rewrite ^/app1/(.*)$ /$1 break: Strips the /app1/ prefix from the request URL before sending it to the backend. This is crucial if your app1 service expects requests at its root (/) instead of /app1/.
• proxy_pass http://app1:80/: Points to your Kubernetes service named app1 (assuming it's listening on port 80). In Kubernetes, services in the same namespace are resolvable via their name directly.
• The proxy_set_header lines ensure the backend gets proper client and host information.

Next, you need to package this config into an Nginx image. Create a Dockerfile:

FROM nginx:alpine
COPY nginx.conf /etc/nginx/nginx.conf

Then build and push the image to a container registry (like Docker Hub, or your private registry):

# Replace <your-registry> and <image-name> with your details
docker build -t <your-registry>/nginx-multipath-proxy:v1 .
docker push <your-registry>/nginx-multipath-proxy:v1

Now deploy this custom Nginx proxy to Kubernetes. Create a file named nginx-proxy.yaml:

apiVersion: apps/v1
kind: Deployment
metadata:
  name: nginx-proxy
spec:
  replicas: 2
  selector:
    matchLabels:
      app: nginx-proxy
  template:
    metadata:
      labels:
        app: nginx-proxy
    spec:
      containers:
      - name: nginx-proxy
        image: <your-registry>/nginx-multipath-proxy:v1
        ports:
        - containerPort: 80
---
apiVersion: v1
kind: Service
metadata:
  name: nginx-proxy-service
spec:
  type: NodePort # Use LoadBalancer if you're on a cloud provider with external IP support
  selector:
    app: nginx-proxy
  ports:
  - protocol: TCP
    port: 80
    targetPort: 80
    nodePort: 30080 # Optional: Specify a port if needed

Apply this to your cluster:

kubectl apply -f nginx-proxy.yaml

• Make sure your app1 and app2 services are running in the same namespace as the Nginx proxy. If they're in a different namespace, update the proxy_pass URLs to use the full service DNS name, e.g., http://app1.other-namespace.svc.cluster.local:80/.
• Test the routing: If using NodePort, access http://<your-node-ip>:30080/app1 and http://<your-node-ip>:30080/app2 to confirm they reach the right services.
• If your backend services don't need the path stripped (e.g., they're already configured to handle /app1), you can remove the rewrite line and adjust proxy_pass to http://app1:80 (without the trailing slash).

内容的提问来源于stack exchange，提问作者Valentin