Jsonnnet配置prometheus-operator, 快速生成yaml资源清单

kube-prometheus

1. 简介

Prometheus 是一套开源的系统监控、报警、时间序列数据库的组合，而 Prometheus Operator 是 CoreOS 开源的一套用于管理在 Kubernetes 集群上的 Prometheus 控制器，它是为了简化在 Kubernetes 上部署、管理和运行 Prometheus 和 Alertmanager 集群。

prometheus-operator与kube-prometheus做对比

prometheus-operator使用kubernetes的CRD自定义资源来简化了prometheus、alertmanager和相关监控组件的部署和配置

kube-prometheus提供了基于prometheus和prometheus-operator的完整集群监控示例配置,包括

部署多个 Prometheus和 Alertmanager 实例, 实现高可用
metrics exporters(如用于收集节点 metrics 的 node_exporters)
将 Prometheus 链接到各种度量端点的目标配置
用于通知集群中潜在问题的示例警报规则

Jsonnet是Google开源的一门配置语言，用于弥补JSON所暴露的短板，它完全兼容JSON，并加入了JSON所没有的一些特性，包括注释、引用、算数运算、条件操作符、数组和对象深入、引入函数、局部变量、继承等，Jsonnet程序被编译为兼容JSON的数据格式。简单来说Jsonnet就是增强版JSON。

kube-prometheus项目内容由Jsonnet编写, 包含了以下组件的基础模板, 除此以外,还包含了默认的dashboard仪表盘和 alerts规则

Prometheus Operator
Highly available Prometheus
Highly available Alertmanager
Prometheus node-exporter
Prometheus Adapter for Kubernetes Metrics APIs
kube-state-metrics
Grafana

2. 安装

安装golang语言和git

wget https://golang.google.cn/dl/go1.13.15.linux-amd64.tar.gz

yum -y install git

解压

tar -xf  go1.13.15.linux-amd64.tar.gz
mv go /opt/software/go1.13.15
ln -s /opt/software/go1.13.15  /opt/software/go

环境变量

cat >> /etc/profile<<EOF
# Golang
SOFT=/opt/software
export GOROOT=$SOFT/go
export GOPATH=$SOFT/gopath
export GO111MODULE=on
export GOPROXY=https://goproxy.io,direct
export PATH=${PATH}:$GOROOT/bin:${GOPATH}/bin
EOF

source /etc/profile

版本测试

$ go version
go version go1.13.15 linux/amd64

安装jsonnet-bundler jsonnet gojsontoyaml

go get github.com/jsonnet-bundler/jsonnet-bundler/cmd/jb
go get github.com/google/go-jsonnet/cmd/jsonnet
go get github.com/brancz/gojsontoyaml

创建目录

mkdir my-kube-prometheus ;  cd my-kube-prometheus

初始化目录并下载kube-prometheus

# 创建初始(空)的 jsonnetfile.json
jb init

$ cat jsonnetfile.json
{
  "version": 1,
  "dependencies": [],
  "legacyImports": true
}

安装kube-prometheus基础依赖

version="main"

# 需要FQ
# 创建 `vendor/ ` & `jsonnetfile.lock.json`，并填写`jsonnetfile.json`
jb install github.com/prometheus-operator/kube-prometheus/jsonnet/kube-prometheus@${version}

# 示例jsonnet文件
wget https://raw.githubusercontent.com/prometheus-operator/kube-prometheus/${version}/example.jsonnet -O example.jsonnet

# build.sh 会根据 jsonnnet文件生成对应的yaml文件
wget https://raw.githubusercontent.com/prometheus-operator/kube-prometheus/${version}/build.sh -O build.sh

执行命令生成资源对应yaml文件

sh build.sh example.jsonnet

执行yaml文件

注意当前环境网络是否能下载国外的镜像仓库
apply和 delete 资源都有顺序要求.

kubectl apply -f manifests/setup/.
kubectl apply -f manifests/.

3.语法说明

(有待补充)

以示例example.json为准

文本一

local kp =
  (import 'kube-prometheus/main.libsonnet') +
  // Uncomment the following imports to enable its patches
  // (import 'kube-prometheus/addons/anti-affinity.libsonnet') +
  // (import 'kube-prometheus/addons/managed-cluster.libsonnet') +

kp为定义的本地变量
import 'kube-prometheus/main.libsonnet' 导入指定文件的内容, 此处为prometheus-operator主要入口文件, 实际路径为 vendor/kube-prometheus/main.libsonnet,

此处为什么可以省略了vendor路径却可以找到对应的文件路径? 是因为build.sh脚本中的命令, -J参数指定了对应的目录路径
```
jsonnet -J vendor -m manifests "${1-example.jsonnet}" | xargs -I{} sh -c 'cat {} | gojsontoyaml > {}.yaml' -- {}
```

文本二

  {
    values+:: {
      common+: {
        namespace: 'monitoring',
      },
    },
  };

+ 在基础配置文本中追加指定内容, 而不覆盖原始.
:: 隐式字段, 不会在jsonnet执行后输出, 可作为内部变量使用
common+:{namespace: 'monitoring',} 根据配置文件定义了全局的命名空间

文本三

{ 'setup/0namespace-namespace': kp.kubePrometheus.namespace } +
{
  ['setup/prometheus-operator-' + name]: kp.prometheusOperator[name]
  for name in std.filter((function(name) name != 'serviceMonitor' && name != 'prometheusRule'), std.objectFields(kp.prometheusOperator))
} +
// serviceMonitor and prometheusRule are separated so that they can be created after the CRDs are ready
{ 'prometheus-operator-serviceMonitor': kp.prometheusOperator.serviceMonitor } +
{ 'prometheus-operator-prometheusRule': kp.prometheusOperator.prometheusRule } +
{ 'kube-prometheus-prometheusRule': kp.kubePrometheus.prometheusRule } +
{ ['alertmanager-' + name]: kp.alertmanager[name] for name in std.objectFields(kp.alertmanager) } +

setup/0namespace-namespace 指定了yaml配置文件指定的路径, (注意, setup目录是由buld.sh脚本生成,而非自动生成)
kp.kubePrometheus.namespace 指定了namespace配置内容
std.filter : 根据文档标准格式为std.filter(func, arr), 返回一个新数组arr，其中包含 func函数返回 true 的所有元素。
std.objectFields: 根据文档标准格式为std.objectFields(o), 返回一个字符串数组，每个元素都是给定对象的一个字段。不包括隐藏字段。
// : 为注释内容, 可在空行注释, 也可以行尾进行注释,

4.应用场景

jsonnnet配置文件中由于过长或重复, 会在文章中选择 `...` 省略大部分内容

Typora暂时没有代码着重显示功能, 所以会在新增内容前后添加`// ******`

4.1. etcd外部服务监听

githup示例

复制证书至当前目录中(也可指定其他路径的证书), 生成yaml后可删除,

# 由kubernetes集群提供证书文件, 包含当前所需文件ca.pem etcd.pem etcd-key.pem
cp /opt/software/kubernetes/certs .

编写jsonnet配置文件 example.jsonnet

local kp =
  (import 'kube-prometheus/main.libsonnet') +
  // Uncomment the following imports to enable its patches
  // (import 'kube-prometheus/addons/anti-affinity.libsonnet') +
  // (import 'kube-prometheus/addons/managed-cluster.libsonnet') +
  // (import 'kube-prometheus/addons/node-ports.libsonnet') +
  // ******
  (import 'kube-prometheus/addons/static-etcd.libsonnet') +
  // ******
  // (import 'kube-prometheus/addons/custom-metrics.libsonnet') +
  // (import 'kube-prometheus/addons/external-metrics.libsonnet') +
  {
    values+:: {
      common+: {
        namespace: 'monitoring',
      }, //common+:
      // ******
      etcd+: {
        ips: ["10.0.10.53","10.0.20.109","10.0.20.108"],
        clientCA: importstr 'certs/ca.pem',
        clientCert: importstr  'certs/etcd.pem',
        clientKey: importstr  'certs/etcd-key.pem',
        insecureSkipVerify: true,
      }, // etcd+:
      // ******
    }, //values+::
  };
...
...
...

涉及到的文件 kube-prometheus/addons/static-etcd.libsonnet github.com/etcd-io/etcd/contrib/mixin/mixin.libsonnet

导入基础模板自带的static-etcd.libsonnet文本文件, 代表新增此功能
ips必须IP地址,否则apply报错
证书文件必须存在或为空文件, 否则报错. certs/ca.pem为``example.jsonnet的相对路径按照模板可得secret yaml文件, 其中metadata.name为 kube-etcd-client-certs`, 如无必要,请勿修改,涉及到许多修改
执行build.sh 会生成文件
- 记录etcd证书的secret文件: prometheus-secretEtcdCerts.yaml
- etcd的service文件:prometheus-serviceEtcd.yaml
- etcd的serviceMonitor文件: prometheus-serviceMonitorEtcd.yaml
- etcd的endponits文件: prometheus-endpointsEtcd.yaml
- 追加etcd secret资源: prometheus-prometheus.yaml

4.2. 根据etcd的mixin.libsonnet文件并没有生成etcd的规则文件和dashboard内容

	查找`github.com/etcd-io/etcd/contrib/mixin/mixin.libsonnet`可得两个信息, `prometheusAlerts+::`  `grafanaDashboards+::` , 这两个配置块根本就没有生成, 在查看模板引用关系, 这两个配置块没有被引用 

(详情可查看文件`kube-prometheus/main.libsonnet`  `kube-prometheus/components/prometheus.libsonnet`  `kube-prometheus/components/grafana.libsonnet`)

	根据官方提供的示例新增 (引用上个示例的配置代码)

...
...
...
  {
    values+:: {
      common+: {
        namespace: 'monitoring',
      }, //common+:
      etcd+: {
        ips: ["10.0.10.53","10.0.20.109","10.0.20.108"],
        clientCA: importstr 'certs/ca.pem',
        clientCert: importstr  'certs/etcd.pem',
        clientKey: importstr  'certs/etcd-key.pem',
        insecureSkipVerify: true,
      }, // etcd+:
      // ******
      grafana+: {
        dashboards+:: {
            // 导入etcd本地json文件 (推荐)
            // 'etcd-dashboard.json' : (import 'etcd-dashboard.json')
            // 导入kube-prometheus自带的etcd dashboard json, 当前示例
            'etcd-dashboard.json': (kp.grafanaDashboards['etcd.json']),
        }, // dashboards+::
      }, // grafana+:
      // ******
    }, //values+::
    // ******
    etcd+: {
      prometheusRule:{
        apiVersion: 'monitoring.coreos.com/v1',
        kind: 'PrometheusRule',
        metadata: {
          name: 'my-prometheus-rule',
          namespace: $.values.common.namespace,
        },
        spec: {
          groups: [ if x['name'] == 'etcd' then  x for x in kp.prometheusAlerts.groups],
        },
      },
    },
    // ******
  };

{ 'setup/0namespace-namespace': kp.kubePrometheus.namespace } +
{
  ['setup/prometheus-operator-' + name]: kp.prometheusOperator[name]
  for name in std.filter((function(name) name != 'serviceMonitor' && name != 'prometheusRule'), std.objectFields(kp.prometheusOperator))
} +
// serviceMonitor and prometheusRule are separated so that they can be created after the CRDs are ready
{ 'prometheus-operator-serviceMonitor': kp.prometheusOperator.serviceMonitor } +
{ 'prometheus-operator-prometheusRule': kp.prometheusOperator.prometheusRule } +
{ 'kube-prometheus-prometheusRule': kp.kubePrometheus.prometheusRule } +
{ ['alertmanager-' + name]: kp.alertmanager[name] for name in std.objectFields(kp.alertmanager) } +
{ ['blackbox-exporter-' + name]: kp.blackboxExporter[name] for name in std.objectFields(kp.blackboxExporter) } +
{ ['grafana-' + name]: kp.grafana[name] for name in std.objectFields(kp.grafana) } +
{ ['kube-state-metrics-' + name]: kp.kubeStateMetrics[name] for name in std.objectFields(kp.kubeStateMetrics) } +
{ ['kubernetes-' + name]: kp.kubernetesControlPlane[name] for name in std.objectFields(kp.kubernetesControlPlane) }
{ ['node-exporter-' + name]: kp.nodeExporter[name] for name in std.objectFields(kp.nodeExporter) } +
{ ['prometheus-' + name]: kp.prometheus[name] for name in std.objectFields(kp.prometheus) } +
{ ['prometheus-adapter-' + name]: kp.prometheusAdapter[name] for name in std.objectFields(kp.prometheusAdapter) } +
// ******
{ ['etcd-' + name]: kp.etcd[name] for name in std.objectFields(kp.etcd) }
// ******

新增三个配置

values+::.grafana 由于kube-prometehus自带了etcd的grafanadashboard模板内容, 示例中直接引用, 但发现此dashboard模板并不好看, 仅做参考.
etcd-dashboard.json为grafana dashboard模板编号3070, 在实际环境中更推荐使用此模板
etcd+: 在etcd块中新增 prometheusRule , 根据其他示例编写, 引用github.com/etcd-io/etcd/contrib/mixin/mixin.libsonnet的prometheusAlerts+::规则. (此处有个问题: 如果kp.prometheusAlerts.groups存在多个, 该如何处理, groups为数组, 该如何指定对应的etcd规则内容 )
指定etcd块文件配置路径, 生成etcd前缀的yaml文件

4.3. etcd的yaml文件都是prometheus前缀, 如何修改成etcd前缀

根据 `kube-prometheus/addons/static-etcd.libsonnet`可得, 所有的配置都在prometheus块中, 导致了 etcd的yaml资源生成都为prometheus前缀.

所需需要修改内容对应的块. 请注意:   prometheus的yaml文件指定了etcd的secret资源, 所以在apply prometheus前需要生成etcd的secret.

修改kube-prometheus/addons/static-etcd.libsonnet文件, 请注意备份, 此处对资源名进行位置变更, 会对内容进行省略处理

(import 'github.com/etcd-io/etcd/contrib/mixin/mixin.libsonnet') + {
  values+:: {
    etcd: {
      ...
    },
  },
  // 此处不变
  prometheus+: {
    secretEtcdCerts: {
      ...
    },
    prometheus+: {
      ...
  },
  // 此处新增了etcd块, 与prometheus区分
  etcd+: {
    serviceEtcd: {
      ...
    },
    endpointsEtcd: {
      ...
    },
    serviceMonitorEtcd: {
      ...
    },
  },
}

根据example.jsonnet文件

{ ['etcd-' + name]: kp.etcd[name] for name in std.objectFields(kp.etcd) }

生成yaml文件

etcd-endpointsEtcd.yaml etcd-prometheusRule.yaml etcd-serviceEtcd.yaml etcd-serviceMonitorEtcd.yaml prometheus-secretEtcdCerts.yaml prometheus-prometheus.yaml的secret内容

4.4. prometheus数据持久化

githup示例

请提前部署nfs-server服务端和nfs-client-provisioner

...
...
...
    // prometheus Start
    prometheus+: {     
        // prometheus配置
        prometheus+:{
            spec+: {
                retention: '30d',
                storage: {
                    volumeClaimTemplate: {
                        apiVersion: 'v1',
                        kind: 'PersistentVolumeClaim',
                        metadata: {
                            name: "prometheus-claim",
                            annotations: {
                                'volume.beta.kubernetes.io/storage-class': "managed-nfs-storage",
                            }, // annotations
                        }, // metadata
                        spec: {
                            accessModes: ['ReadWriteMany'],
                            resources: { requests: { storage: '10Gi' } },
                        }, // spec
                    }, // volumeClaimTemplate
                }, // storage:
            }, // spec+
        }, // prometheus+:
    }, // prometheus+:
    // prometheus End
...
...
...

retention: 数据保留时间默认为24h , 如有没有指定该值, 则会通过prometheus-operator传递参数--storage.tsdb.retention=24h给prometheus-k8s
resources.requests.storage: prometheus-k8s存储空间大小
查看prometheus-prometheus.yaml 可以看到资源信息

  storage:
    volumeClaimTemplate:
      apiVersion: v1
      kind: PersistentVolumeClaim
      metadata:
        annotations:
          volume.beta.kubernetes.io/storage-class: managed-nfs-storage
        name: prometheus-claim
      spec:
        accessModes:
        - ReadWriteMany
        resources:
          requests:
            storage: 10Gi

4.5. prometheus添加外部暴露地址, 添加前缀路径

默认:  prometheus的访问地址为 http://192.168.1.100:9090/ ,  当前使用ingress暴露prometheus服务,却又不想新增二级域名给prometheus.

match: Host(test.k8s.com) && PathPrefix(/prometheus)

在配置ingressroute资源发现访问prometheus服务最终会重定向到 http://test.k8s.com/graph 链接去.  与配置不符合, 导致404报错

在trafik找不到对应的解决办法, 那么就只能更改后端的路径, 给路径添加前缀.  修改后会影响prometheus-adapter访问, 自身prometheus监听, grafana默认数据源

...
...
...
  {
    values+:: {
        common+: {
            namespace: 'monitoring',
        }, //common+::
        prometheusAdapter+: {
            prometheusURL: 'http://prometheus-' + $.values.prometheus.name + '.' + $.values.common.namespace + '.svc.cluster.local:9090/prometheus',
        },
    }, // values+::

    // prometheus Start
    prometheus+: {
        serviceMonitor+:{
            spec+:{
                endpoints:[
                    {"path": "/prometheus/metrics"},
                ],
            },
        },
        // prometheus配置
        prometheus+:{
            spec+: {
                retention: '30d',
                externalUrl: "http://test.k8s.com/prometheus", //外部暴露地址
                routePrefix: "/prometheus", // 添加前缀路径
            } // spec+
        } // prometheus+:
    }, // prometheus+:
    // prometheus End
...
...
...

prometheusAdapter 对应服务 prometheus-adapter需要访问k8s集群内部prometheus指标信息, 而此处需求因变更prometheus实际访问路径,需要变更参数值. 参考 kube-prometheus/main.libsonnet 文件中 prometheusAdapter 的 prometheusURL
prometheus.serviceMonitor 需要监听 prometheus提供的指标信息, 同样的情况, 指定了metrics的路径(默认为: /metrics)
prometheus.spec.externalUrl: 该参数定义了外部暴露地址, 现阶段看到 alertmanager使用
prometheus.spec.routePrefix: 该参数定义了prometheus实际的访问路径, 添加前缀地址/prometheus

4.6. prometheus通过ingressroute暴露服务

前提: 部署traefik服务并监听monitor命名空间

当前配置中定义了ingressroute函数, 通过传递参数形式实现配置.

githup示例中, 定义的是ingress资源, 且直接定义了ingress的部分.

而这里定义的是IngressRoute资源, 将ingress的部分写入到对应服务代码部分中, 让ingress生成的文件以服务名称前缀命名

// ******
local ingressroute(name, namespace, entrypoint, route) = {
  apiVersion: 'traefik.containo.us/v1alpha1',
  kind: 'IngressRoute',
  metadata: {
    name: name,
    namespace: namespace,
  },
  spec: { 
    entryPoints: entrypoint,
    routes: route
  },
};
// ******
local kp =
  (import 'kube-prometheus/main.libsonnet') +

  {
    values+:: {
        common+: {
            namespace: 'monitoring',
        }, //common+::
    }, // values+::
    // ******
    prometheus+:{
        ingressroute+: ingressroute(
            'prometheus',
            $.values.common.namespace,
            [ 'web' ],
            [{
              match: 'Host(`test.k8s.com`) && PathPrefix(`/prometheus`)',
              kind: 'Rule',
              services: [{
                name: 'prometheus-k8s',
                port: 9090,
              }],
            }],
        ), // ingressroute
    }, // prometheus+
  // ******
  };
...
...
...

4.7. grafana添加前缀路径

问题与prometehus一样, 通过ingressroute暴露服务后,  访问中会出现重定向地址, 导致404报错

match: Host(test.k8s.com) && PathPrefix(/grafana)

修改grafana配置

    // grafana Start
    grafana+:: {
        _config+:: {
            env+: [
                {"name": "GF_SERVER_ROOT_URL", "value": "http://localhost:3000/grafana/"},
                {"name": "GF_SERVER_SERVE_FROM_SUB_PATH", "value": "true"},
                {"name": "GF_LOG_LEVEL", "value": "debug"},
            ], // env
        }, //  _config+:
    }, // _grafana+:
    // grafana End

参考文件内容: github.com/brancz/kubernetes-grafana/grafana/grafana.libsonnet

GF_SERVER_ROOT_URL: 定义grafana的访问路径地址, 后缀添加grafana, 对应参数root_url
GF_SERVER_SERVE_FROM_SUB_PATH: 从root_url设置中指定实际访问路径对应参数serve_from_sub_path

4.8. grafana修改默认数据源

参考模板文件: github.com/brancz/kubernetes-grafana/grafana/grafana.libsonnet

由prometheus新增了前缀路径 导致grafana生成的默认数据源不生效, 而且在grafana页面修改默认数据源信息也不生效, 而新添加的数据源没问题.

    // grafana Start
    grafana+: {
        _config+:: {
            datasources: [{
                name: 'Prometheus', // 该名称对应dashboard模板文件的数据源指向,注意大小写
                type: 'prometheus',
                access: 'proxy',
                orgId: 1,
                url: 'http://prometheus-k8s.monitoring.svc:9090/prometheus',
                version: 1,
                editable: false,
            }], //datasources+::
        }, //  _config+:
    }, // _grafana+:
    // grafana End

查看yaml

$ cat manifests/grafana-dashboardDatasources.yaml | grep "datasources.yaml" | cut -d " " -f 4 | base64 -d
{
    "apiVersion": 1,
    "datasources": [
        {
            "access": "proxy",
            "editable": false,
            "name": "prometheus",
            "orgId": 1,
            "type": "prometheus",
            "url": "http://prometheus-k8s.monitoring.svc:9090/prometheus",
            "version": 1
        }
    ]
}

url 默认为: 'http://' + $._config.prometheus.serviceName + '.' + $._config.namespace + '.svc:9090', 在example.jsonnet无法正确的引用, 所以此处写死url地址.

4.9. grafana启动前添加自定义dashboard

[githup示例](https://github.com/prometheus-operator/kube-prometheus/blob/main/examples/grafana-additional-rendered-dashboard-example.jsonnet)

从grafana官网上下载dashboard模板json文件至目录中,  json示例编号为13105  : [kubernetes-dashboard.json](https://grafana.com/api/dashboards/13105/revisions/5/download)

wget https://grafana.com/api/dashboards/13105/revisions/5/download -O kubernetes-dashboard.json

  {
    values+:: {
        common+: {
            namespace: 'monitoring',
        }, //common+::
        grafana+: {
            // grafana-dashboardDefinitions.yaml
            // 需要注意dashboard文件数据源指向是否对应正确拼写的数据源名称
            dashboards+:: {
                'kubernetes-dashboard.json' : (import 'kubernetes-dashboard.json'),
            }, // dashboards+::
        }, // grafana+
    }, // values+::

$ grep "kubernetes-dashboard" manifests/grafana-dashboardDefinitions.yaml
    kubernetes-dashboard.json: |-
    name: grafana-dashboard-kubernetes-dashboard

在grafana的UI界面中可以看到对应的dashboard

4.10. grafana持久化

根据实际情况进行修改

前提: 搭建nfs-server 和 nfs-client-provisioner

github.com/brancz/kubernetes-grafana/grafana/grafana.libsonnet 查看三个参数 storageVolumeName storageVolumeName storageVolumeMount , 可以看到都是添加了local关键字, 导致无法从example.jsonnet进行修改

直接修改模板配置文件
vim github.com/brancz/kubernetes-grafana/grafana/grafana.libsonnet

# 原始
local storageVolumeName = 'grafana-storage';
local storageVolume = { name: storageVolumeName, emptyDir: {} };
local storageVolumeMount = { name: storageVolumeName, mountPath: '/var/lib/grafana', readOnly: false };

# 修改后
local storageVolumeName = 'grafana-storage';
local storageVolumeClaim = 'grafana-claim';
local storageVolume =  { name: storageVolumeName, persistentVolumeClaim: { claimName: storageVolumeClaim } };
local storageVolumeMount = { name: storageVolumeName, mountPath: '/var/lib/grafana', readOnly: false };

storageVolumeClaim grafana-claim 为 nfs的资源PersistentVolumeClaim定义的名称

    grafana+: {
        PersistentVolumeClaim: {
            apiVersion: 'v1',
            kind: 'PersistentVolumeClaim',
            metadata: {
                name: 'grafana-claim',
                namespace: $.values.common.namespace,
                annotations: {
                    'volume.beta.kubernetes.io/storage-class' : 'managed-nfs-storage',
                }, // annotations
            }, // metadata
            spec: {
                accessModes: ['ReadWriteMany'],
                resources: { requests: { storage: '10G'},},
            }, // spec
        }, // PersistentVolumeClaim:
    }, // grafana+:

4.11. grafana通过ingressroute暴露服务

前提: 部署traefik服务并监听monitor命名空间

当前配置中定义了ingressroute函数, 通过传递参数形式实现配置.

githup示例中, 定义的是ingress资源, 且直接定义了ingress的部分.

而这里定义的是IngressRoute资源, 将ingress的部分写入到对应服务代码部分中, 让ingress生成的文件以服务名称前缀命名

// ******
local ingressroute(name, namespace, entrypoint, route) = {
  apiVersion: 'traefik.containo.us/v1alpha1',
  kind: 'IngressRoute',
  metadata: {
    name: name,
    namespace: namespace,
  },
  spec: { 
    entryPoints: entrypoint,
    routes: route
  },
};
// ******
local kp =
  (import 'kube-prometheus/main.libsonnet') +
  {
    values+:: {
        common+: {
            namespace: 'monitoring',
        }, //common+::
    }, // values+::
    // ******
    grafana+:{
        ingressroute+: ingressroute(
            'grafana',
            $.values.common.namespace,
            [ 'web' ],
            [{
              match: 'Host(`test.k8s.com`) && PathPrefix(`/grafana`)',
              kind: 'Rule',
              services: [{
                name: 'grafana',
                port: 3000,
              }],
            }],
        ), // ingressroute
    }, // grafana+
  // ******
  };
...
...
...

4.12. alertmanager钉钉告警配置

alertmanger githup示例

alertmanager当前不支持直接往钉钉告警信息, 需要通过dingtalk服务来进行告警信息, 告警自定义模板可参考alertmanager告警模板定义模板文档

(个人编写的jsonnet文件, 后续有待调整, 会将参数变为可定义的值, template_configmap配置里面相关的URL需要根据实际情况进行修改, )

编写 dingtalk.libsonnet, 修改 url 的 token , secret , @的人的手机号码(可多人)

// 这是作用于钉钉机器人的告警 有待进行下一步的调整
// 请修改 config_configmap资源里面的url地址  secret  mobiles  text:  @your-@-phone
{
    dingtalk: {
        // service
        service: {
            apiVersion: 'v1',
            kind: 'Service',
            metadata: { 
                labels: { 
                    app: 'alertmanager-webhook-dingtalk' 
                },
                name: 'alertmanager-webhook-dingtalk',
                namespace: $.values.common.namespace,
            },
            spec: { 
                selector: { 
                    app: 'alertmanager-webhook-dingtalk' 
                },
                ports: [ 
                    { name: 'http', port: 8060, targetPort: 'http' } 
                    ]
            },
        },
        // deployment
        deployment: {
            apiVersion: 'apps/v1',
            kind: 'Deployment',
            metadata: {
                labels: {
                    app: 'alertmanager-webhook-dingtalk' 
                }, // label
                name: 'alertmanager-webhook-dingtalk',
                namespace:  $.values.common.namespace,
            }, // metadata
            spec: {
                selector: {
                    matchLabels: {
                        app: 'alertmanager-webhook-dingtalk' 
                    }, // matchLabels
                }, // selector
                template: {
                    metadata: {
                        labels: {
                            app: 'alertmanager-webhook-dingtalk' 
                        }, // labels
                    }, // metadata
                    spec: { 
                        containers: [{ 
                            name: 'alertmanager-webhook-dingtalk',
                            image: 'timonwong/prometheus-webhook-dingtalk',
                            args: [ 
                                '--web.listen-address=:8060',
                                '--config.file=/config/config.yaml',
                                '--web.enable-ui',
                                '--web.enable-lifecycle',
                                '--log.level=info'
                            ], // args
                            volumeMounts: [
                                { 
                                    name: 'dingtalk-config',
                                    mountPath: '/config/config.yaml',
                                    subPath: 'config.yaml' 
                                },
                                {
                                    name: 'dingtalk-template',
                                    mountPath: '/config/template.tmpl',
                                    subPath: 'template.tmpl' 
                                }
                            ], // volumeMounts
                            resources: { limits: { cpu: '100m', memory: '100Mi' } },
                            ports: [ { name: 'http', containerPort: 8060 } ] 
                        }], // containers
                        volumes: [
                            { name: 'dingtalk-config',configMap: { name: 'dingtalk-config' } },
                            { name: 'dingtalk-template',configMap: { name: 'dingtalk-template' } } 
                        ] // volumes
                    }, // spec
                }, // template
            }, // spec
        }, // deployment
        // template_configmap
        template_configmap:{
            apiVersion: 'v1',
            kind: 'ConfigMap',
            metadata: {
                name: 'dingtalk-template',
                namespace:  $.values.common.namespace,
            },
            data: {
                'template.tmpl': |||
                    {{/* Alert List Begin */}}
                    {{ define "example.__text_alert_list" }}{{ range . }}

                    **{{ .Annotations.message }}**
                    
                    [Prometheus]({{ .GeneratorURL }}) | [Alertmanager](https://alertmanager.example.com/#/alerts) | [Grafana](https://grafana.example.com/dashboards)

                    **\[{{ .Labels.severity | upper }}\]** {{ .Annotations.summary }}

                    **Description:** {{ .Annotations.description }}

                    {{ range .Labels.SortedPairs }}> - {{ .Name }}: {{ .Value | markdown | html }}
                    {{ end }}
                    {{ end }}{{ end }}
                    {{/* Alert List End */}}

                    {{/* Message Title Begin */}}
                    
                    {{ define "example.title" }}{{ template "__subject" . }}{{ end }}
                    {{/* Message Title End */}}

                    {{/* Message Content Begin */}}
                    {{ define "example.content" }}
                    ### [{{ index .GroupLabels "alertname" }}](https://example.app.opsgenie.com/alert/list)

                    {{ if gt (len .Alerts.Firing) 0 -}}
                    {{ template "example.__text_alert_list" .Alerts.Firing }}
                    {{- end }}

                    {{ if gt (len .Alerts.Resolved) 0 -}}
                    {{ template "example.__text_alert_list" .Alerts.Resolved }}
                    {{- end }}
                    {{- end }}
                    {{/* Message Content End */}}
                |||,
            },
        },
        // config_configmap
        config_configmap: {
            apiVersion: 'v1',
            kind: 'ConfigMap',
            metadata: {
                name: 'dingtalk-template',
                namespace:  $.values.common.namespace,
            },
            data: {
                'config.yaml': |||
                    templates:
                    - /config/template.tmpl
                    targets:
                      webhook1: &target_base
                        url: https://oapi.dingtalk.com/robot/send?access_token=xxx
                        # secret for signature
                        secret: xxx
                        message:
                          title: '{{ template "example.title" . }}'
                          text: '{{ template "example.content" . }}'
                      webhook2:
                        <<: *target_base
                        mention:
                          mobiles: ['phone1','phone2']
                        message:
                          text: |
                            @phone1 @phone2
                            
                            {{ template "example.content" . }}
                |||,
            },
        },
    },
}

编写alertmanager.yaml文件

alertmanager.yaml名称不要更改, 否则无法覆盖alertmanager的对应文件

global:
  resolve_timeout: 5m
route:
  group_by: ["severity",'alertname']
  receiver: 'webhook1'
  routes:
  - match_re:
      severity: warning
    receiver: webhook2
receivers:
- name: webhook1
  webhook_configs:
  - &dingtalk_config
    url: 'http://alertmanager-webhook-dingtalk.monitoring.svc.cluster.local:8060/dingtalk/webhook1/send'
- name: webhook2
  webhook_configs:
  - <<: *dingtalk_config
    url: 'http://alertmanager-webhook-dingtalk.monitoring.svc.cluster.local:8060/dingtalk/webhook2/send'

编写example.jsonnet

在测试中发现不配置configSecret参数,, 上面的配置信息无法在pod中覆盖原配置.

local kp =
  (import 'kube-prometheus/main.libsonnet') +
  // 新增
  (import 'dingtalk.libsonnet') +
  {
    values+:: {
        common+: {
            namespace: 'monitoring',
        }, //common+::
         alertmanager+: {
            config: config: importstr 'alertmanager.yaml',
         }, // alertmanager
    }, // values+::
    alertmanager+: {
        alertmanager+: {
            configSecret: 'alertmanager-main',
        },
    }, 
  };
...
...
{ ['prometheus/prometheus-' + name]: kp.prometheus[name] for name in std.objectFields(kp.prometheus) } +
{ ['prometheus-adapter/prometheus-adapter-' + name]: kp.prometheusAdapter[name] for name in std.objectFields(kp.prometheusAdapter) }+
// 新增
{ ['dingtalk-' + name]: kp.dingtalk[name] for name in std.objectFields(kp.dingtalk) }

4.13. alertmanger通过ingressroute暴露服务

super与python的super意义一样,调用父类

// ******
local ingressroute(name, namespace, entrypoint, route) = {
  apiVersion: 'traefik.containo.us/v1alpha1',
  kind: 'IngressRoute',
  metadata: {
    name: name,
    namespace: namespace,
  },
  spec: { 
    entryPoints: entrypoint,
    routes: route
  },
};
// ******
local kp =
  (import 'kube-prometheus/main.libsonnet') +
  {
    values+:: {
        common+: {
            namespace: 'monitoring',
        }, //common+::
    }, // values+::
    prometheus+: {
        prometheus+:{
            spec+: {
                alerting+:{ alertmanagers: [super.alertmanagers[0] + { pathPrefix: '/alertmanager'} ]},
            },
        },
    }, // prometheus
    alertmanager+:{
        alertmanager+: {
            routePrefix: '/alertmanager'
        },
        ingressroute+: ingressroute(
            'alertmanager',
            $.values.common.namespace,
            [ 'web' ],
            [{
              match: 'Host(`test.k8s.com`) && PathPrefix(`/alertmanager`)',
              kind: 'Rule',
              services: [{
                name: kp.alertmanager.service.metadata.name,
                port: kp.alertmanager.service.spec.ports[0]['port']
              }],
            }],
        ), // ingressroute
    }, // alertmanager+
  // ******
  };
...
...
...

4.14 将服务以目录形式存放文件

在example.jsonnet路径中添加前缀目录即可如 ['grafana-' + name] 更改为 ['grafana/grafana-' + name]

但会出现一个问题, yaml文件都存放在各自目录需要手动一个目录一个目录执行

修改build.sh脚本, 按照需求进行变更

# 原始
mkdir -p manifests/setup
# 修改
mkdir -p manifests/{setup,prometheus-operator,prometheus,alertmanager,blackbox-exporter,grafana,kube-state-metrics,kubernetes,node-exporter,prometheus-adapter}

修改example.json

{ 'setup/0namespace-namespace': kp.kubePrometheus.namespace } +
{
  ['setup/prometheus-operator-' + name]: kp.prometheusOperator[name]
  for name in std.filter((function(name) name != 'serviceMonitor' && name != 'prometheusRule'), std.objectFields(kp.prometheusOperator))
} +
// serviceMonitor and prometheusRule are separated so that they can be created after the CRDs are ready
{ 'prometheus-operator/prometheus-operator-serviceMonitor': kp.prometheusOperator.serviceMonitor } +
{ 'prometheus-operator/prometheus-operator-prometheusRule': kp.prometheusOperator.prometheusRule } +
{ 'prometheus/kube-prometheus-prometheusRule': kp.kubePrometheus.prometheusRule } +
{ ['alertmanager/alertmanager-' + name]: kp.alertmanager[name] for name in std.objectFields(kp.alertmanager) } +
{ ['blackbox-exporter/blackbox-exporter-' + name]: kp.blackboxExporter[name] for name in std.objectFields(kp.blackboxExporter) } +
{ ['grafana/grafana-' + name]: kp.grafana[name] for name in std.objectFields(kp.grafana) } +
{ ['kube-state-metrics/kube-state-metrics-' + name]: kp.kubeStateMetrics[name] for name in std.objectFields(kp.kubeStateMetrics) } +
{ ['kubernetes/kubernetes-' + name]: kp.kubernetesControlPlane[name] for name in std.objectFields(kp.kubernetesControlPlane) }
{ ['node-exporter/node-exporter-' + name]: kp.nodeExporter[name] for name in std.objectFields(kp.nodeExporter) } +
{ ['prometheus/prometheus-' + name]: kp.prometheus[name] for name in std.objectFields(kp.prometheus) } +
{ ['prometheus-adapter/prometheus-adapter-' + name]: kp.prometheusAdapter[name] for name in std.objectFields(kp.prometheusAdapter) } +

执行yamll ,setup目录执行的优先级最高, 因为需要先生成CRD自定义资源

kubectl apply -f manifests/setup
kubectl apply -f manifests/prometheus
kubectl apply -f manifests/prometheus-operator
kubectl apply -f manifests/grafana
kubectl apply -f manifests/prometheus-adapter
kubectl apply -f manifests/kubernetes
kubectl apply -f manifests/blackbox-exporter
kubectl apply -f manifests/kube-state-metrics
kubectl apply -f manifests/node-exporter
kubectl apply -f manifests/alertmanager

删除yaml

kubectl delete -f manifests/grafana
kubectl delete -f manifests/prometheus-adapter
kubectl delete -f manifests/kubernetes
kubectl delete -f manifests/blackbox-exporter
kubectl delete -f manifests/kube-state-metrics
kubectl delete -f manifests/node-exporter
kubectl delete -f manifests/alertmanager
kubectl delete -f manifests/prometheus-operator
kubectl delete -f manifests/prometheus
kubectl delete -f manifests/setup

5. 问题

问题一: grafana默认prometheus数据源信息不正确且修改无效

通过prometheus-oprator安装方式安装, 对prometheus访问路径添加前缀后, grafana默认prometheus数据源变更也无法访问,

该问题由于grafana挂载了datasources.yaml配置文件, 写死了prometheus url地址, 无论如何修改, 始终都是该url地址

所以需要找到datasources.yaml的定义位置.

kube-prometheus/manifests/grafana-dashboardDatasources.yaml

该配置文件是一组加密的配置文件,该如何修改呢

apiVersion: v1
data:
  datasources.yaml: ewogICAgImFwaVZlcnNpb24iOiAxLAogICAgImRhdGFzb3VyY2VzIjogWwogICAgICAgIHsKICAgICAgICAgICAgImFjY2VzcyI6ICJwcm94eSIsCiAgICAgICAgICAgICJlZGl0YWJsZSI6IGZhbHNlLAogICAgICAgICAgICAibmFtZSI6ICJwcm9tZXRoZXVzIiwKICAgICAgICAgICAgIm9yZ0lkIjogMSwKICAgICAgICAgICAgInR5cGUiOiAicHJvbWV0aGV1cyIsCiAgICAgICAgICAgICJ1cmwiOiAiaHR0cDovL3Byb21ldGhldXMtazhzLm1vbml0b3Jpbmcuc3ZjOjkwOTAiLAogICAgICAgICAgICAidmVyc2lvbiI6IDEKICAgICAgICB9CiAgICBdCn0=
kind: Secret
metadata:
  labels:
    app.kubernetes.io/component: grafana
    app.kubernetes.io/name: grafana
    app.kubernetes.io/part-of: kube-prometheus
    app.kubernetes.io/version: 8.0.3
  name: grafana-datasources
  namespace: monitoring
type: Opaque

解压

echo -n "ewogICAgImFwaVZlcnNpb24iOiAxLAogICAgImRhdGFzb3VyY2VzIjogWwogICAgICAgIHsKICAgICAgICAgICAgImFjY2VzcyI6ICJwcm94eSIsCiAgICAgICAgICAgICJlZGl0YWJsZSI6IGZhbHNlLAogICAgICAgICAgICAibmFtZSI6ICJwcm9tZXRoZXVzIiwKICAgICAgICAgICAgIm9yZ0lkIjogMSwKICAgICAgICAgICAgInR5cGUiOiAicHJvbWV0aGV1cyIsCiAgICAgICAgICAgICJ1cmwiOiAiaHR0cDovL3Byb21ldGhldXMtazhzLm1vbml0b3Jpbmcuc3ZjOjkwOTAiLAogICAgICAgICAgICAidmVyc2lvbiI6IDEKICAgICAgICB9CiAgICBdCn0="| base64 --decode

{
    "apiVersion": 1,
    "datasources": [
        {
            "access": "proxy",
            "editable": false,
            "name": "prometheus",
            "orgId": 1,
            "type": "prometheus",
            "url": "http://prometheus-k8s.monitoring.svc:9090",
            "version": 1
        }
    ]
}

生成临时文件tmp.json 修改url

{
    "apiVersion": 1,
    "datasources": [
        {
            "access": "proxy",
            "editable": false,
            "name": "prometheus",
            "orgId": 1,
            "type": "prometheus",
            "url": "http://prometheus-k8s.monitoring.svc:9090/prometheus",
            "version": 1
        }
    ]
}

加密

base64 < tmp.json | tr -d '\n'

ewogICAgImFwaVZlcnNpb24iOiAxLAogICAgImRhdGFzb3VyY2VzIjogWwogICAgICAgIHsKICAgICAgICAgICAgImFjY2VzcyI6ICJwcm94eSIsCiAgICAgICAgICAgICJlZGl0YWJsZSI6IGZhbHNlLAogICAgICAgICAgICAibmFtZSI6ICJwcm9tZXRoZXVzIiwKICAgICAgICAgICAgIm9yZ0lkIjogMSwKICAgICAgICAgICAgInR5cGUiOiAicHJvbWV0aGV1cyIsCiAgICAgICAgICAgICJ1cmwiOiAiaHR0cDovL3Byb21ldGhldXMtazhzLm1vbml0b3Jpbmcuc3ZjOjkwOTAvcHJvbWV0aGV1cyIsCiAgICAgICAgICAgICJ2ZXJzaW9uIjogMQogICAgICAgIH0KICAgIF0KfQo=

修改配置文件对应位置, 重新apply和restart相关资源

问题二: alertmanger高可用无法组成集群

level=warn ts=2021-08-19T09:51:05.495Z caller=cluster.go:251 component=cluster msg="failed to join cluster" err="3 errors occurred:\n\t* Failed to resolve alertmanager-main-0.alertmanager-operated:9094: lookup alertmanager-main-0.alertmanager-operated on 10.254.0.2:53: no such host\n\t* Failed to resolve alertmanager-main-1.alertmanager-operated:9094: lookup alertmanager-main-1.alertmanager-operated on 10.254.0.2:53: no such host\n\t* Failed to resolve alertmanager-main-2.alertmanager-operated:9094: lookup alertmanager-main-2.alertmanager-operated on 10.254.0.2:53: no such host\n\n"

一般都为coredns问题,请重启coredns服务.
测试: 在busybox容器中使用ping命令成功, 使用nslookup失败. 在alertmanager-main-0中使用nsllokup失败. 排除alertmanager问题. 重启coredns后继续测试, nslookup没问题,且alertmanger日志无报错, 成功使用

6. 参考链接

[kube-prometheus]

kube-prometheus

kuber-prometheus文章

nfs-client-provisioner

[prometheus-operator]

prometheus官方

grafana dashbaord官方

prometheus-operator个人文章

dingtalk githup

[jsonnet]

jsonnnet官方

jsonnet语法参考

jsonnet stackoverflow

jsonnet为什么从JSON构建而不是YAML 官方

7. 完整实例

根据应用场景所有功能编写的example.jsonnet文件, 仅做参考, 需要根据实际情况修改

// 准备文件  
// 服务器可FQ获取镜像资源
// etcd : ca.pem etcd.pem etcd-key.pem  修改etcd
// prometheus:  dnsUrl
// grafana:  dnsurl
// dingtalk template:  token secret
// nfs-provisioner 服务 (后续可以自行编写jsonnet)\
// 修改build.sh mkdir生成目录加入服务前缀

local ingressroute(name, namespace, entrypoint, route) = {
  apiVersion: 'traefik.containo.us/v1alpha1',
  kind: 'IngressRoute',
  metadata: {
    name: name,
    namespace: namespace,
  },
  spec: { 
    entryPoints: entrypoint,
    routes: route
  },
};

local kp =
  (import 'kube-prometheus/main.libsonnet') +
  (import 'kube-prometheus/addons/static-etcd.libsonnet') +
  (import 'dingtalk.libsonnet') +
  {
    values+:: {
        common+: {
            namespace: 'monitoring',
        }, //common+:

        // etcd外部服务监听
        etcd+: {
            ips: ["10.0.10.53","10.0.20.109","10.0.20.108"],
            clientCA: importstr 'ca.pem',
            clientCert: importstr  'etcd.pem',
            clientKey: importstr  'etcd-key.pem',
            insecureSkipVerify: true,
        }, // etcd+:

        prometheus+: {
            prometheus+:{
                spec+: {
                    alerting+:{ alertmanagers: super.alertmanagers[0] + { pathPrefix: '/alertmanager'} },
                },
            },
        }, // prometheus

        prometheusAdapter+: {
            prometheusURL: 'http://prometheus-' + $.values.prometheus.name + '.' + $.values.common.namespace + '.svc.cluster.local:9090/prometheus',
            hostNetwork: true
        },

        // 根据etcd的mixin.libsonnet文件并没有生成etcd的规则文件和dashboard内容
        grafana+: {
            dashboards+:: {
                'kubernetes-dashboard.json' : (import 'kubernetes-dashboard.json'),
                // 导入etcd本地json文件 (推荐)
                'etcd-dashboard.json' : (import 'etcd-dashboard.json')
                // 导入kube-prometheus自带的etcd dashboard json, 当前示例
                // 'etcd-dashboard.json': (kp.grafanaDashboards['etcd.json']),
            }, // dashboards+::
        }, // grafana+:

       alertmanager+: {
            config: importstr 'alertmanager.yaml',
       }, // alertmanager
    }, //values+::
    etcd+: {
        prometheusRule:{
            apiVersion: 'monitoring.coreos.com/v1',
            kind: 'PrometheusRule',
            metadata: {
                name: 'etcd-rule',
                namespace: $.values.common.namespace,
            }, // metadata
            spec: {
                groups: [ if x['name'] == 'etcd' then  x for x in kp.prometheusAlerts.groups],
            }, // spec
        }, // prometheusRule
    }, // etcd

    // prometheus Start
    prometheus+: {
        serviceMonitor+:{
            spec+:{
                endpoints:[
                    {"path": "/prometheus/metrics"},
                ],
            },
        },
        // prometheus配置
        prometheus+:{
            spec+: {
                retention: '30d',
                externalUrl: "http://test.k8s-host.com/prometheus", //外部暴露地址
                routePrefix: "/prometheus", // 添加前缀路径
                storage: {
                    volumeClaimTemplate: {
                        apiVersion: 'v1',
                        kind: 'PersistentVolumeClaim',
                        metadata: {
                            name: "prometheus-claim",
                            annotations: {
                                'volume.beta.kubernetes.io/storage-class': "managed-nfs-storage",
                            }, // annotations
                        }, // metadata
                        spec: {
                            accessModes: ['ReadWriteMany'],
                            resources: { requests: { storage: '10Gi' } },
                        }, // spec
                    }, // volumeClaimTemplate
                }, // storage:
            }, // spec+
        }, // prometheus+:

        ingressroute+: ingressroute(
            'prometheus',
            $.values.common.namespace,
            [ 'web' ],
            [{
              match: 'Host(`test.k8s-host.com`) && PathPrefix(`/prometheus`)',
              kind: 'Rule',
              services: [{
                name: 'prometheus-k8s',
                port: 9090,
              }],
            }],
        ), // ingressroute
    }, // prometheus

    grafana+: {
        _config+:: {
            env+: [
                {"name": "GF_SERVER_ROOT_URL", "value": "http://localhost:3000/grafana/"},
                {"name": "GF_SERVER_SERVE_FROM_SUB_PATH", "value": "true"},
                {"name": "GF_LOG_LEVEL", "value": "debug"},
            ], // env
            datasources: [{
                name: 'Prometheus', // 该名称对应dashboard模板文件的数据源指向,注意大小写
                type: 'prometheus',
                access: 'proxy',
                orgId: 1 ,
                url: 'http://prometheus-k8s.monitoring.svc:9090/prometheus',
                version: 1,
                editable: false,
            }], //datasources+::
        }, //  _config+:
        PersistentVolumeClaim: {
            apiVersion: 'v1',
            kind: 'PersistentVolumeClaim',
            metadata: {
                name: 'grafana-claim', // 与 grafana.libsonnet的'storageVolumeClaim' 一致
                namespace: $.values.common.namespace,
                annotations: {
                    'volume.beta.kubernetes.io/storage-class' : 'managed-nfs-storage',
                }, // annotations
            }, // metadata
            spec: {
                accessModes: ['ReadWriteMany'],
                resources: { requests: { storage: '10G'},},
            }, // spec
        }, // PersistentVolumeClaim:
        ingressroute+: ingressroute(
            'grafana',
            $.values.common.namespace,
            [ 'web' ],
            [{
              match: 'Host(`test.k8s-host.com`) && PathPrefix(`/grafana`)',
              kind: 'Rule',
              services: [{
                name: 'grafana',
                port: 3000,
              }],
            }],
        ), // ingressroute
    }, // _grafana+:
    alertmanager+: {
        alertmanager+: {
            routePrefix: '/alertmanager'
            configSecret: 'alertmanager-main',
        },
        ingressroute+: ingressroute(
            'alertmanager',
            $.values.common.namespace,
            [ 'web' ],
            [{
              match: 'Host(`test.k8s.com`) && PathPrefix(`/alertmanager`)',
              kind: 'Rule',
              services: [{
                name: kp.alertmanager.service.metadata.name,
                port: kp.alertmanager.service.spec.ports[0]['port']
              }],
            }],
        ), // ingressroute
    },
  };

{ 'setup/0namespace-namespace': kp.kubePrometheus.namespace } +
{
  ['setup/prometheus-operator-' + name]: kp.prometheusOperator[name]
  for name in std.filter((function(name) name != 'serviceMonitor' && name != 'prometheusRule'), std.objectFields(kp.prometheusOperator))
} +
// serviceMonitor and prometheusRule are separated so that they can be created after the CRDs are ready
{ 'prometheus-operator/prometheus-operator-serviceMonitor': kp.prometheusOperator.serviceMonitor } +
{ 'prometheus-operator/prometheus-operator-prometheusRule': kp.prometheusOperator.prometheusRule } +
{ 'prometheus/kube-prometheus-prometheusRule': kp.kubePrometheus.prometheusRule } +
{ ['alertmanager/alertmanager-' + name]: kp.alertmanager[name] for name in std.objectFields(kp.alertmanager) } +
{ ['blackbox-exporter/blackbox-exporter-' + name]: kp.blackboxExporter[name] for name in std.objectFields(kp.blackboxExporter) } +
{ ['grafana/grafana-' + name]: kp.grafana[name] for name in std.objectFields(kp.grafana) } +
{ ['kube-state-metrics/kube-state-metrics-' + name]: kp.kubeStateMetrics[name] for name in std.objectFields(kp.kubeStateMetrics) } +
{ ['kubernetes/kubernetes-' + name]: kp.kubernetesControlPlane[name] for name in std.objectFields(kp.kubernetesControlPlane) }
{ ['node-exporter/node-exporter-' + name]: kp.nodeExporter[name] for name in std.objectFields(kp.nodeExporter) } +
{ ['prometheus/prometheus-' + name]: kp.prometheus[name] for name in std.objectFields(kp.prometheus) } +
{ ['prometheus-adapter/prometheus-adapter-' + name]: kp.prometheusAdapter[name] for name in std.objectFields(kp.prometheusAdapter) } +
{ ['etcd/etcd-' + name]: kp.etcd[name] for name in std.objectFields(kp.etcd) } +
{ ['dingtalk/dingtalk-' + name]: kp.dingtalk[name] for name in std.objectFields(kp.dingtalk) }

posted @ 2021-08-18 17:14 风吹蛋生丶阅读(696) 评论(0) 编辑收藏举报

刷新页面返回顶部

风吹蛋生丶

Jsonnnet配置prometheus-operator, 快速生成yaml资源清单

kube-prometheus

1. 简介

2. 安装

3.语法说明

4.应用场景

4.1. etcd外部服务监听

4.2. 根据etcd的mixin.libsonnet文件并没有生成etcd的规则文件和dashboard内容

4.3. etcd的yaml文件都是prometheus前缀, 如何修改成etcd前缀

4.4. prometheus数据持久化

4.5. prometheus添加外部暴露地址, 添加前缀路径

4.6. prometheus通过ingressroute暴露服务

4.7. grafana添加前缀路径

4.8. grafana修改默认数据源

4.9. grafana启动前添加自定义dashboard

4.10. grafana持久化

4.11. grafana通过ingressroute暴露服务

4.12. alertmanager钉钉告警配置

4.13. alertmanger通过ingressroute暴露服务

4.14 将服务以目录形式存放文件

5. 问题

问题一: grafana默认prometheus数据源信息不正确且修改无效

问题二: alertmanger高可用无法组成集群

6. 参考链接

7. 完整实例

公告