Prometheus-Operator 安装

记录自己安装 Prometheus Operator 的过程

一、Prometheus Operator 介绍

kubernetes的监控系统Prometheus 应该都比较了解, 简述以下几点吧

1. Prometheus 简介

组件

Prometheus 由多个组件组成，但是其中许多组件是可选的：

Prometheus Server：用于抓取指标、存储时间序列数据
exporter：暴露指标让任务来抓
pushgateway：push 的方式将指标数据推送到该网关
alertmanager：处理报警的报警组件
adhoc：用于数据查询
架构
下图是 Prometheus 官方提供的架构及其一些相关的生态系统组件：

2. Operator 简介

Operator 是 CoreOS 推出的旨在简化复杂有状态应用管理的框架，它是一个感知应用状态的控制器，通过扩展 Kubernetes API 来自动创建、管理和配置应用实例。

你可以在 OperatorHub.io 上查看 Kubernetes 社区推荐的一些 Operator 范例。

operator 架构图

prometheus-operator-架构图

二、安装依赖

安装Helm

1. 下载Helm

Hlem版本我使用的是当前2版本中的最新版 2.16.7，因为官方建议使用2.14以上版本, 不然会有CRD相关问题, 具体见Github prometheus-operator

wget https://storage.googleapis.com/kubernetes-helm/helm-v2.16.7-linux-amd64.tar.gz
tar zxvf helm-v2.16.7-linux-amd64.tar.gz
mv linux-amd64/helm linux-amd64/tiller /usr/local/bin/
helm version

2. 创建RBAC

创建文件rbac-tiller.yaml , 内容为下

apiVersion: v1
kind: ServiceAccount
metadata:
  name: tiller
  namespace: kube-system
---
apiVersion: rbac.authorization.k8s.io/v1beta1
kind: ClusterRoleBinding
metadata:
  name: tiller
roleRef:
  apiGroup: rbac.authorization.k8s.io
  kind: ClusterRole
  name: cluster-admin
subjects:
  - kind: ServiceAccount
    name: tiller
    namespace: kube-system

创建RBAC

kubectl apply -f rbac-tiller.yaml

2. 初始化Helm

因为默认下载gcr.io仓库的镜像, 由于墙的原因下载失败, 所以我下载后传到了我司仓库

helm init --service-account tiller --tiller-image registry.cn-beijing.aliyuncs.com/roobo/tiller:v2.16.7
$ helm version
Client: &version.Version{SemVer:"v2.16.7", GitCommit:"5f2584fd3d35552c4af26036f0c464191287986b", GitTreeState:"clean"}
Server: &version.Version{SemVer:"v2.16.7", GitCommit:"5f2584fd3d35552c4af26036f0c464191287986b", GitTreeState:"clean"}

三、安装Prometheus-Operator

1. 创建Namespace

(将相关PODs都创建到此NS下)

kubelet create ns monitoring

2. 安装 prometheus-operator

helm install --namespace monitoring  --name prometheus-operator stable/prometheus-operator

3. 查看相关PODs

[root@k8smaster-01 ~]# kubectl -n monitoring get po
NAME                                                     READY     STATUS    RESTARTS   AGE
alertmanager-prometheus-operator-alertmanager-0           2/2       Running   0          53m
prometheus-operator-grafana-69bfccc949-h9s7x              2/2       Running   0          53m
prometheus-operator-kube-state-metrics-7ddcbdb744-xzh9w   1/1       Running   0          53m
prometheus-operator-operator-6d4f47dc49-9g9jr             2/2       Running   0          53m
prometheus-operator-prometheus-node-exporter-h9c2p        1/1       Running   0          53m
prometheus-operator-prometheus-node-exporter-jw2hn        1/1       Running   0          53m
prometheus-operator-prometheus-node-exporter-mqq4p        1/1       Running   0          53m
prometheus-operator-prometheus-node-exporter-zxcg5        1/1       Running   0          53m
prometheus-prometheus-operator-prometheus-0               3/3       Running   1          53m

[root@k8smaster-01 ~]# kubectl -n monitoring get svc
NAME                                           TYPE        CLUSTER-IP      EXTERNAL-IP   PORT(S)                      AGE
alertmanager-operated                          ClusterIP   None            <none>        9093/TCP,9094/TCP,9094/UDP   53m
prometheus-operated                            ClusterIP   None            <none>        9090/TCP                     53m
prometheus-operator-alertmanager               ClusterIP   10.254.13.40    <none>        9093/TCP                     54m
prometheus-operator-grafana                    ClusterIP   10.254.0.159    <none>        80/TCP                       54m
prometheus-operator-kube-state-metrics         ClusterIP   10.254.43.177   <none>        8080/TCP                     54m
prometheus-operator-operator                   ClusterIP   10.254.38.46    <none>        8080/TCP,443/TCP             54m
prometheus-operator-prometheus                 ClusterIP   10.254.27.218   <none>        9090/TCP                     54m
prometheus-operator-prometheus-node-exporter   ClusterIP   10.254.60.8     <none>        9100/TCP                     54m

[root@k8smaster-01 ~]# kubectl get crd
NAME                                    CREATED AT
alertmanagers.monitoring.coreos.com     2020-05-10T06:38:44Z
podmonitors.monitoring.coreos.com       2020-05-10T06:38:51Z
prometheuses.monitoring.coreos.com      2020-05-10T06:38:56Z
prometheusrules.monitoring.coreos.com   2020-05-10T06:39:02Z
servicemonitors.monitoring.coreos.com   2020-05-10T06:39:07Z
thanosrulers.monitoring.coreos.com      2020-05-10T06:39:12Z

4. 创建Ingress

默认情况下Grafana并不能直接访问, 可以将svc改为NodePort方式或者创建Ingress 通过域名的方式访问到, 这里以Ingress举例

Yaml文件

(将$DOMAIN修改为自己的域名)

apiVersion: extensions/v1beta1
kind: Ingress
metadata:
  name: ingress-prometheus-operator-grafana
  namespace: monitoring
spec:
  rules:
  - host: $DOMAIN
    http:
      paths:
      - path: /
        backend:
          serviceName: prometheus-operator-grafana
          servicePort: 80