使用traefik作为kubernetes的ingress

说明

关于kubernetes的服务暴露方式以及traefik的原理篇这里不作详细说明。traefik的原理可以参考官方文档:https://docs.traefik.io/ ,而关于kubernetes的服务暴露方式以及将traefik作为kubernetes的ingress使用的基本原理也可以参考这篇文档:https://mritd.me/2016/12/06/try-traefik-on-kubernetes/ 。这篇文档主要阐述将traefik作为kubernetes的ingress的实际操作,包括在traefik中启用https,将traefik的日志输出为json并通过filebeat收集,以及通过prometheus监控traefik。

部署

创建一个独立的命名空间

kubectl create ns ingress

配置rbac授权

我这里kubernetes使用1.8且启用了rbac授权。官方的traefik-rbac.yaml文件可以参考:https://raw.githubusercontent.com/containous/traefik/master/examples/k8s/traefik-rbac.yaml ,但在我实际测试中,还是有问题。我这里直接采用了一个偷懒的做法,直接给了最大权限。 下面的配置文件中,先创建了一个serviceaccount名称为traefik-ingress-controller,然后直接为这个serviceaccount授予了cluster-admin的权限

traefik.rbac.yaml示例如下:

apiVersion: v1
kind: ServiceAccount
metadata:
  labels:
    k8s-app: traefik-ingress-controller
  name: traefik-ingress-controller
  namespace: ingress
---
apiVersion: rbac.authorization.k8s.io/v1
kind: ClusterRoleBinding
metadata:
  name: traefik-ingress-controller
  labels:
    k8s-app: traefik-ingress-controller
roleRef:
  apiGroup: rbac.authorization.k8s.io
  kind: ClusterRole
  name: cluster-admin
subjects:
- kind: ServiceAccount
  name: traefik-ingress-controller
  namespace: ingress

配置secret

我们将traefik作为ingress时,需要其同时处理https请求,所以需要为traefik配置证书文件。在这里将证书文件和key文件放入到secret中。我这里两个文件的名称分别是breezey.top.crt和breezey.top.key。创建secret方法如下:

kubectl create secret tls breezey-ingress-secret --key breezey.top.key --cert breezey.top.crt -n ingress

创建一个configmap用于存放traefik的配置文件

traefik.configmap.yaml配置如下:

apiVersion: v1
kind: ConfigMap
metadata:
  name: traefik-conf
  namespace: ingress
data:
  traefik.toml: |
    defaultEntryPoints = ["http", "https"]
    [entryPoints]
      [entryPoints.http]
      address = ":80"
      [entryPoints.https]
      address = ":443"
        [entryPoints.https.tls]
          [[entryPoints.https.tls.certificates]]
          certFile = "/keys/breezey.top.crt"
          keyFile = "/keys/breezey.top.key"
    #traefikLogsFile = "log/traefik.log"
    [accessLog]
    filePath = "/logs/traefik.access.log"
    format = "json"

需要说明的是,我这里同时支持http和https,并没有做http强制跳转到https。如果需要http强制跳转到https,可以参考如下配置:

defaultEntryPoints = ["http", "https"]
[entryPoints]
  [entryPoints.http]
  address = ":80"
    [entryPoints.http.redirect]
    entryPoint = "https"
  [entryPoints.https]
  address = ":443"
    [entryPoints.https.tls]
      [[entryPoints.https.tls.certificates]]
      certFile = "/keys/breezey.top.crt"
      keyFile = "/keys/breezey.top.key"
#traefikLogsFile = "log/traefik.log"
[accessLog]
filePath = "/logs/traefik.access.log"
format = "json"

另外,在上面的配置中,我启动了traefik的访问日志,并将格式指定为json,这是为了方便后面直接通过filebeat作收集。这里也需要说明下,traefik的两种日志,一种是traefik服务自身的日志,另一种是访问日志。我这里只启用了访问日志。默认traefik的服务日志会通过容器的标准输出打印。

配置traefik的deployment文件

在这里我采用nodeSelecter的方式将traefik固定在指定的两个节点上,如果采用deployment的方式部署,需要在选定的节点上先打好标签。
traefik.dm.yaml配置如下:

apiVersion: extensions/v1beta1
kind: Deployment
metadata:
  name: traefik-ingress-lb
  namespace: ingress
  labels:
    k8s-app: traefik-ingress-lb
spec:
  strategy:
    type: RollingUpdate
    rollingUpdate:
      maxUnavailable: 1
      maxSurge: 0
  replicas: 2
  selector: 
    matchLabels:
      k8s-app: traefik-ingress-lb
  template:
    metadata:
      labels:
        k8s-app: traefik-ingress-lb
      annotations:
        prometheus.io/scrape: "true"
        prometheus.io/port: "8580"
    spec:
      terminationGracePeriodSeconds: 60
      hostNetwork: true
      restartPolicy: Always
      volumes:
      - name: traefik-config
        configMap:
          name: traefik-conf
      - name: traefik-key
        secret:
          secretName: breezey-ingress-secret
      - name: traefik-log
        hostPath:
          path: /mnt/srvlogs
      - name: localtime
        hostPath:
          path: /etc/localtime
      containers:
      - image: dyhub.douyucdn.cn/library/traefik:v1.4.3
        name: traefik-ingress-lb
#        resources:
#          limits:
#            cpu: 200m
#            memory: 30Mi
#          requests:
#            cpu: 100m
#            memory: 20Mi
        securityContext:
          privileged: true
        ports:
        - name: http
          containerPort: 80
        - name: https
          containerPort: 443
        - name: admin
          containerPort: 8580
        volumeMounts:
        - mountPath: "/config"
          name: "traefik-config"
        - mountPath: "/logs"
          name: "traefik-log"
        - mountPath: "/keys"
          name: "traefik-key"
        - mountPath: "/etc/localtime"
          name: "localtime"
        args:
        - --configfile=/config/traefik.toml
        - --web
        - --web.address=:8580
        - --kubernetes
        - --web.metrics.prometheus
      serviceAccountName: traefik-ingress-controller
      nodeSelector:
        proxy: "true"
        ingress: "traefik"

需要说明的是,通过volumeMounts的方式将访问日志以及https证书及密钥进行了相关挂载。然后在启动参数中,指定--web.metrics.prometheus参数以暴露metrics供prometheus收集,另外指定了管理端口为8580。

最后创建上面配置的所有资源:

kubectl create -f ./

到此,traefik在kubernetes上部署完成。

配置服务

这里是一个将traefik-ui作为traefik的代理后端的示例。

先创建一个service文件traefik-ui.svc.yaml示例如下:

apiVersion: v1
kind: Service
metadata:
  name: traefik-web-ui
  namespace: ingress
spec:
  clusterIP: None
  selector:
    k8s-app: traefik-ingress-lb
  ports:
  - name: web
    port: 8580
    targetPort: 8580

再创建一个traefik-ui.ingress.yaml文件如下:

apiVersion: extensions/v1beta1
kind: Ingress
metadata:
  annotations:
    kubernetes.io/ingress.class: traefik
  name: traefik-web-ui
  namespace: ingress
spec:
  tls:
#  - hosts:
#    - traefik-ui.breezey.top
  - secretName: breezey-ingress-secret
  rules:
  - host: traefik-ui.breezey.top
    http:
      paths:
      - path: /
        backend:
          serviceName: traefik-web-ui
          servicePort: web

创建相关资源:

kubectl create -f ./traefik-ui.svc.yaml
kubectl create -f ./traefik-ui.ingress.yaml

配置好dns解析,即可通过traefik-ui.breezey.top来访问traefik-ui服务,http和https同时支持,且不会强制跳转。

通过prometheus监控traefik

在启动traefik时,使用了--web.metrics.prometheus选项,这时只需要将traefik服务的ip及admin端口添加至prometheus的配置文件中即可。如下:

  - job_name: 'traefik'
    static_configs:
    - targets: ['192.168.0.101:8580','192.168.0.102:8580']

重启prometheus:

systemctl restart prometheus

在本次示例中,prometheus未运行于kubernetes中,而是独立部署

在prometheus的status的target中查看:
prometheus-target

在grafana中添加dashboard,这里采用的官方的id为2870的dashboard,添加后效果如下:
grafana-traefik

posted @ 2018-05-30 22:26  breezey  阅读(3318)  评论(0编辑  收藏  举报