Question

我正在尝试将Nginx与nginx-prometheus-exporter一起部署。以下是我的部署

apiVersion: apps/v1beta1
kind: Deployment
metadata:
  namespace: local
  name: nginx
spec:
  replicas: 1
  template:
    metadata:
      labels:
        app: nginx
    spec:
      containers:
      - name: nginx
        image: nginx
        imagePullPolicy: Always
        ports:
        - name: nginx-port
          containerPort: 80
          protocol: TCP
        livenessProbe:
          httpGet:
            path: /
            port: nginx-port
            httpHeaders:
              - name: Host
                value: KubernetesLivenessProbe
          initialDelaySeconds: 10
          timeoutSeconds: 1
          periodSeconds: 15
        readinessProbe:
          httpGet:
            path: /
            port: nginx-port
            httpHeaders:
              - name: Host
                value: KubernetesLivenessProbe
          initialDelaySeconds: 3
          timeoutSeconds: 1
          periodSeconds: 8
        volumeMounts:
        - mountPath: /etc/nginx/conf.d/ # mount nginx-conf volume to /etc/nginx
          readOnly: true
          name: frontend-conf
      - name: nginx-exporter
        image: nginx/nginx-prometheus-exporter:0.3.0
        resources:
          requests:
            cpu: 100m
            memory: 100Mi
        ports:
        - name: nginx-ex-port
          containerPort: 9113
        args:
          - -nginx.scrape-uri=http://nginx.local.svc.cluster.local:80/stub_status
      volumes:
      - name: frontend-conf
        configMap:
          name: frontend-conf # place ConfigMap `nginx-conf` on /etc/nginx
          items:
            - key: frontend.local.conf
              path: frontend.local.conf

单个容器pod可以工作，但是当我添加nginx-prometheus-exporter sidecar容器时失败。

mackbook: xyz$ kubectl logs nginx-6dbbdb64fb-8drmc -c nginx-exporter
2019/05/14 20:17:48 Starting NGINX Prometheus Exporter Version=0.3.0 GitCommit=6570275
2019/05/14 20:17:53 Could not create Nginx Client: Failed to create NginxClient: failed to get http://nginx.local.svc.cluster.local:80/stub_status: Get http://nginx.local.svc.cluster.local:80/stub_status: net/http: request canceled while waiting for connection (Client.Timeout exceeded while awaiting headers)

还能够从另一个运行在不同名称空间的容器中curl http://nginx.local.svc.cluster.local。

任何人都知道指定-nginx.scrape-uri的正确方法是什么？

从Nginx容器

root@nginx-6fd866c4d7-t8wwf:/# curl localhost/stub_status
<html>
<head><title>404 Not Found</title></head>
<body>
<h1>404 Not Found</h1>
<ul>
<li>Code: NoSuchKey</li>
<li>Message: The specified key does not exist.</li>
<li>Key: beta/stub_status</li>
<li>RequestId: 3319E5A424C7BA31</li>
<li>HostId: zVv5zuAyzqbB2Gw3w1DavEJwgq2sbgI8RPMf7RmPsNQoh/9ftxCmiSwyj/j6q/K3lxEEeUy6aZM=</li>
</ul>
<hr/>
</body>
</html>
root@nginx-6fd866c4d7-t8wwf:/#

Answer 1

nginx容器正在侦听端口80，并且您将Sidecar配置为访问8080端口。而且，您可以在“生产”名称空间中部署它，但是使用“测试”来访问nginx。

尝试使用-nginx.scrape-uri=http://nginx.production.svc.cluster.local:80/stub_status。

顺便说一句，您必须知道Pod中的容器共享一个IP地址和端口空间，并且可以通过localhost相互找到。

因此您可以使用-nginx.scrape-uri=http://localhost/stub_status。

nginx-prometheus-exporter部署失败

1 个答案: