前言

Prometheus是一个用于监控和告警的开源系统。一开始由Soundcloud开发，后来在2016年，它迁移到CNCF并且称为Kubernetes之后最流行的项目之一。从整个Linux服务器到stand-alone web服务器、数据库服务或一个单独的进程，它都能监控。在Prometheus术语中，它所监控的事物称为目标（Target）。每个目标单元被称为指标（metric）。它以设置好的时间间隔通过http抓取目标，以收集指标并将数据放置在其时序数据库（Time Series Database）中。你可以使用PromQL查询语言查询相关target的指标。

本文中，我们将一步一步展示如何：

安装Prometheus（使用prometheus-operator Helm chart）以基于自定义事件进行监控/告警
创建和配置自定义告警规则，它将会在满足条件时发出告警
集成Alertmanager以处理由客户端应用程序（在本例中为Prometheus server）发送的告警
将Alertmanager与发送告警通知的邮件账户集成。

理解Prometheus及其抽象概念

从下图我们将看到所有组成Prometheus生态的组件：

Prometheus生态中的所有组件（来源：）

以下是与本文相关的术语，大家可以快速了解：

Prometheus Server ：在时序数据库中抓取和存储指标的主要组件

抓取：一种拉取方法以获取指标。它通常以10-60秒的时间间隔抓取。

Target ：检索数据的server客户端
服务发现 ：启用Prometheus，使其能够识别它需要监控的应用程序并在动态环境中拉取指标
Alert Manager ：负责处理警报的组件（包括silencing、inhibition、聚合告警信息，并通过邮件、PagerDuty、Slack等方式发送告警通知）。
数据可视化 ：抓取的数据存储在本地存储中，并使用PromQL直接查询，或通过Grafana dashboard查看。

理解Prometheus Operator

根据Prometheus Operator的项目所有者CoreOS称，Prometheus Operator可以配置原生Kubernetes并且可以管理和操作Prometheus和Alertmanager集群。

该Operator引入了以下Kubernetes自定义资源定义（CRDs）：Prometheus、ServiceMonitor、PrometheusRule和Alertmanager。如果你想了解更多内容可以访问链接：

https://github.com/coreos/prometheus-operator/blob/master/Documentation/design.md

在我们的演示中，我们将使用PrometheusRule来定义自定义规则。

首先，我们需要使用 stable/prometheus-operator Helm chart来安装Prometheus Operator，下载链接：

https://github.com/helm/charts/tree/master/stable/prometheus-operator

默认安装程序将会部署以下组件：prometheus-operator、prometheus、alertmanager、node-exporter、kube-state-metrics以及grafana。默认状态下，Prometheus将会抓取Kubernetes的主要组件：kube-apiserver、kube-controller-manager以及etcd。

安装Prometheus软件

前期准备

要顺利执行此次demo，你需要准备以下内容：

一个Google Cloud Platform账号（免费套餐即可）。其他任意云也可以
Rancher v2.3.5（发布文章时的最新版本）
运行在GKE（版本1.15.9-gke.12.）上的Kubernetes集群（使用EKS或AKS也可以）
在计算机上安装好Helm binary

启动一个Rancher实例

直接按照这一直观的入门指南进行操作即可：

https://rancher.com/quick-start

使用Rancher部署一个GKE集群

使用Rancher来设置和配置你的Kubernetes集群：

https://rancher.com/docs/rancher/v2.x/en/cluster-provisioning/hosted-kubernetes-clusters/gke/

部署完成后，并且为kubeconfig文件配置了适当的credential和端点信息，就可以使用kubectl指向该特定集群。

部署Prometheus 软件

首先，检查一下我们所运行的Helm版本

$ helm version
version.BuildInfo{Version:"v3.1.2", GitCommit:"d878d4d45863e42fd5cff6743294a11d28a9abce", GitTreeState:"clean", GoVersion:"go1.13.8"}

当我们使用Helm 3时，我们需要添加一个stable 镜像仓库，因为默认状态下不会设置该仓库。

$ helm repo add stable https://kubernetes-charts.storage.googleapis.com
"stable" has been added to your repositories

$ helm repo update
Hang tight while we grab the latest from your chart repositories...
...Successfully got an update from the "stable" chart repository
Update Complete. ⎈ Happy Helming!⎈

$ helm repo list
NAME    URL
stable  https://kubernetes-charts.storage.googleapis.com

Helm配置完成后，我们可以开始安装prometheus-operator

$ kubectl create namespace monitoring
namespace/monitoring created

$ helm install --namespace monitoring demo stable/prometheus-operator
manifest_sorter.go:192: info: skipping unknown hook: "crd-install"
manifest_sorter.go:192: info: skipping unknown hook: "crd-install"
manifest_sorter.go:192: info: skipping unknown hook: "crd-install"
manifest_sorter.go:192: info: skipping unknown hook: "crd-install"
manifest_sorter.go:192: info: skipping unknown hook: "crd-install"
manifest_sorter.go:192: info: skipping unknown hook: "crd-install"
NAME: demo
LAST DEPLOYED: Sat Mar 14 09:40:35 2020
NAMESPACE: monitoring
STATUS: deployed
REVISION: 1
NOTES:
The Prometheus Operator has been installed. Check its status by running:
  kubectl --namespace monitoring get pods -l "release=demo"

Visit https://github.com/coreos/prometheus-operator for instructions on how
to create & configure Alertmanager and Prometheus instances using the Operator.

规则

除了监控之外，Prometheus还让我们创建触发告警的规则。这些规则基于Prometheus的表达式语言。只要满足条件，就会触发告警并将其发送到Alertmanager。之后，我们会看到规则的具体形式。

我们回到demo。Helm完成部署之后，我们可以检查已经创建了什么pod：

$ kubectl -n monitoring get pods
NAME                                                   READY   STATUS    RESTARTS   AGE
alertmanager-demo-prometheus-operator-alertmanager-0   2/2     Running   0          61s
demo-grafana-5576fbf669-9l57b                          3/3     Running   0          72s
demo-kube-state-metrics-67bf64b7f4-4786k               1/1     Running   0          72s
demo-prometheus-node-exporter-ll8zx                    1/1     Running   0          72s
demo-prometheus-node-exporter-nqnr6                    1/1     Running   0          72s
demo-prometheus-node-exporter-sdndf                    1/1     Running   0          72s
demo-prometheus-operator-operator-b9c9b5457-db9dj      2/2     Running   0          72s
prometheus-demo-prometheus-operator-prometheus-0       3/3     Running   1          50s

为了从web浏览器中访问Prometheus和Alertmanager，我们需要使用port转发。

由于本例中使用的是GCP实例，并且所有的kubectl命令都从该实例运行，因此我们使用实例的外部IP地址访问资源。

$ kubectl port-forward --address 0.0.0.0 -n monitoring prometheus-demo-prometheus-operator-prometheus-0 9090  >/dev/null 2>&1 &

$ kubectl port-forward --address 0.0.0.0 -n monitoring alertmanager-demo-prometheus-operator-alertmanager-0 9093  >/dev/null 2>&1 &

“Alert”选项卡向我们展示了所有当前正在运行/已配置的告警。也可以通过查询名称为prometheusrules的CRD从CLI进行检查：

$ kubectl -n monitoring get prometheusrules
NAME                                                            AGE
demo-prometheus-operator-alertmanager.rules                     3m21s
demo-prometheus-operator-etcd                                   3m21s
demo-prometheus-operator-general.rules                          3m21s
demo-prometheus-operator-k8s.rules                              3m21s
demo-prometheus-operator-kube-apiserver-error                   3m21s
demo-prometheus-operator-kube-apiserver.rules                   3m21s
demo-prometheus-operator-kube-prometheus-node-recording.rules   3m21s
demo-prometheus-operator-kube-scheduler.rules                   3m21s
demo-prometheus-operator-kubernetes-absent                      3m21s
demo-prometheus-operator-kubernetes-apps                        3m21s
demo-prometheus-operator-kubernetes-resources                   3m21s
demo-prometheus-operator-kubernetes-storage                     3m21s
demo-prometheus-operator-kubernetes-system                      3m21s
demo-prometheus-operator-kubernetes-system-apiserver            3m21s
demo-prometheus-operator-kubernetes-system-controller-manager   3m21s
demo-prometheus-operator-kubernetes-system-kubelet              3m21s
demo-prometheus-operator-kubernetes-system-scheduler            3m21s
demo-prometheus-operator-node-exporter                          3m21s
demo-prometheus-operator-node-exporter.rules                    3m21s
demo-prometheus-operator-node-network                           3m21s
demo-prometheus-operator-node-time                              3m21s
demo-prometheus-operator-node.rules                             3m21s
demo-prometheus-operator-prometheus                             3m21s
demo-prometheus-operator-prometheus-operator                    3m21s

我们也可以检查位于prometheus容器中prometheus-operator Pod中的物理文件。

$ kubectl -n monitoring exec -it prometheus-demo-prometheus-operator-prometheus-0 -- /bin/sh
Defaulting container name to prometheus.
Use 'kubectl describe pod/prometheus-demo-prometheus-operator-prometheus-0 -n monitoring' to see all of the containers in this pod.

在容器中，我们可以检查规则的存储路径：

/prometheus $ ls /etc/prometheus/rules/prometheus-demo-prometheus-operator-prometheus-rulefiles-0/
monitoring-demo-prometheus-operator-alertmanager.rules.yaml                    monitoring-demo-prometheus-operator-kubernetes-system-apiserver.yaml
monitoring-demo-prometheus-operator-etcd.yaml                                  monitoring-demo-prometheus-operator-kubernetes-system-controller-manager.yaml
monitoring-demo-prometheus-operator-general.rules.yaml                         monitoring-demo-prometheus-operator-kubernetes-system-kubelet.yaml
monitoring-demo-prometheus-operator-k8s.rules.yaml                             monitoring-demo-prometheus-operator-kubernetes-system-scheduler.yaml
monitoring-demo-prometheus-operator-kube-apiserver-error.yaml                  monitoring-demo-prometheus-operator-kubernetes-system.yaml
monitoring-demo-prometheus-operator-kube-apiserver.rules.yaml                  monitoring-demo-prometheus-operator-node-exporter.rules.yaml
monitoring-demo-prometheus-operator-kube-prometheus-node-recording.rules.yaml  monitoring-demo-prometheus-operator-node-exporter.yaml
monitoring-demo-prometheus-operator-kube-scheduler.rules.yaml                  monitoring-demo-prometheus-operator-node-network.yaml
monitoring-demo-prometheus-operator-kubernetes-absent.yaml                     monitoring-demo-prometheus-operator-node-time.yaml
monitoring-demo-prometheus-operator-kubernetes-apps.yaml                       monitoring-demo-prometheus-operator-node.rules.yaml
monitoring-demo-prometheus-operator-kubernetes-resources.yaml                  monitoring-demo-prometheus-operator-prometheus-operator.yaml
monitoring-demo-prometheus-operator-kubernetes-storage.yaml                    monitoring-demo-prometheus-operator-prometheus.yaml

为了详细了解如何将这些规则加载到Prometheus中，请检查Pod的详细信息。我们可以看到用于prometheus容器的配置文件是etc/prometheus/config_out/prometheus.env.yaml。该配置文件向我们展示了文件的位置或重新检查yaml的频率设置。

$ kubectl -n monitoring describe pod prometheus-demo-prometheus-operator-prometheus-0

完整命令输出如下：

Name:           prometheus-demo-prometheus-operator-prometheus-0
Namespace:      monitoring
Priority:       0
Node:           gke-c-7dkls-default-0-c6ca178a-gmcq/10.132.0.15
Start Time:     Wed, 11 Mar 2020 18:06:47 +0000
Labels:         app=prometheus
                controller-revision-hash=prometheus-demo-prometheus-operator-prometheus-5ccbbd8578
                prometheus=demo-prometheus-operator-prometheus
                statefulset.kubernetes.io/pod-name=prometheus-demo-prometheus-operator-prometheus-0
Annotations:    <none>
Status:         Running
IP:             10.40.0.7
IPs:            <none>
Controlled By:  StatefulSet/prometheus-demo-prometheus-operator-prometheus
Containers:
  prometheus:
    Container ID:  docker://360db8a9f1cce8d72edd81fcdf8c03fe75992e6c2c59198b89807aa0ce03454c
    Image:         quay.io/prometheus/prometheus:v2.15.2
    Image ID:      docker-pullable://quay.io/prometheus/prometheus@sha256:914525123cf76a15a6aaeac069fcb445ce8fb125113d1bc5b15854bc1e8b6353
    Port:          9090/TCP
    Host Port:     0/TCP
    Args:
      --web.console.templates=/etc/prometheus/consoles
      --web.console.libraries=/etc/prometheus/console_libraries
      --config.file=/etc/prometheus/config_out/prometheus.env.yaml
      --storage.tsdb.path=/prometheus
      --storage.tsdb.retention.time=10d
      --web.enable-lifecycle
      --storage.tsdb.no-lockfile
      --web.external-url=http://demo-prometheus-operator-prometheus.monitoring:9090
      --web.route-prefix=/
    State:       Running
      Started:   Wed, 11 Mar 2020 18:07:07 +0000
    Last State:  Terminated
      Reason:    Error
      Message:    caller=main.go:648 msg="Starting TSDB ..."
level=info ts=2020-03-11T18:07:02.185Z caller=web.go:506 component=web msg="Start listening for connections" address=0.0.0.0:9090
level=info ts=2020-03-11T18:07:02.192Z caller=head.go:584 component=tsdb msg="replaying WAL, this may take awhile"
level=info ts=2020-03-11T18:07:02.192Z caller=head.go:632 component=tsdb msg="WAL segment loaded" segment=0 maxSegment=0
level=info ts=2020-03-11T18:07:02.194Z caller=main.go:663 fs_type=EXT4_SUPER_MAGIC
level=info ts=2020-03-11T18:07:02.194Z caller=main.go:664 msg="TSDB started"
level=info ts=2020-03-11T18:07:02.194Z caller=main.go:734 msg="Loading configuration file" filename=/etc/prometheus/config_out/prometheus.env.yaml
level=info ts=2020-03-11T18:07:02.194Z caller=main.go:517 msg="Stopping scrape discovery manager..."
level=info ts=2020-03-11T18:07:02.194Z caller=main.go:531 msg="Stopping notify discovery manager..."
level=info ts=2020-03-11T18:07:02.194Z caller=main.go:553 msg="Stopping scrape manager..."
level=info ts=2020-03-11T18:07:02.194Z caller=manager.go:814 component="rule manager" msg="Stopping rule manager..."
level=info ts=2020-03-11T18:07:02.194Z caller=manager.go:820 component="rule manager" msg="Rule manager stopped"
level=info ts=2020-03-11T18:07:02.194Z caller=main.go:513 msg="Scrape discovery manager stopped"
level=info ts=2020-03-11T18:07:02.194Z caller=main.go:527 msg="Notify discovery manager stopped"
level=info ts=2020-03-11T18:07:02.194Z caller=main.go:547 msg="Scrape manager stopped"
level=info ts=2020-03-11T18:07:02.197Z caller=notifier.go:598 component=notifier msg="Stopping notification manager..."
level=info ts=2020-03-11T18:07:02.197Z caller=main.go:718 msg="Notifier manager stopped"
level=error ts=2020-03-11T18:07:02.197Z caller=main.go:727 err="error loading config from \"/etc/prometheus/config_out/prometheus.env.yaml\": couldn't load configuration (--config.file=\"/etc/prometheus/config_out/prometheus.env.yaml\"): open /etc/prometheus/config_out/prometheus.env.yaml: no such file or directory"

      Exit Code:    1
      Started:      Wed, 11 Mar 2020 18:07:02 +0000
      Finished:     Wed, 11 Mar 2020 18:07:02 +0000
    Ready:          True
    Restart Count:  1
    Liveness:       http-get http://:web/-/healthy delay=0s timeout=3s period=5s #success=1 #failure=6
    Readiness:      http-get http://:web/-/ready delay=0s timeout=3s period=5s #success=1 #failure=120
    Environment:    <none>
    Mounts:
      /etc/prometheus/certs from tls-assets (ro)
      /etc/prometheus/config_out from config-out (ro)
      /etc/prometheus/rules/prometheus-demo-prometheus-operator-prometheus-rulefiles-0 from prometheus-demo-prometheus-operator-prometheus-rulefiles-0 (rw)
      /prometheus from prometheus-demo-prometheus-operator-prometheus-db (rw)
      /var/run/secrets/kubernetes.io/serviceaccount from demo-prometheus-operator-prometheus-token-jvbrr (ro)
  prometheus-config-reloader:
    Container ID:  docker://de27cdad7067ebd5154c61b918401b2544299c161850daf3e317311d2d17af3d
    Image:         quay.io/coreos/prometheus-config-reloader:v0.37.0
    Image ID:      docker-pullable://quay.io/coreos/prometheus-config-reloader@sha256:5e870e7a99d55a5ccf086063efd3263445a63732bc4c04b05cf8b664f4d0246e
    Port:          <none>
    Host Port:     <none>
    Command:
      /bin/prometheus-config-reloader
    Args:
      --log-format=logfmt
      --reload-url=http://127.0.0.1:9090/-/reload
      --config-file=/etc/prometheus/config/prometheus.yaml.gz
      --config-envsubst-file=/etc/prometheus/config_out/prometheus.env.yaml
    State:          Running
      Started:      Wed, 11 Mar 2020 18:07:04 +0000
    Ready:          True
    Restart Count:  0
    Limits:
      cpu:     100m
      memory:  25Mi
    Requests:
      cpu:     100m
      memory:  25Mi
    Environment:
      POD_NAME:  prometheus-demo-prometheus-operator-prometheus-0 (v1:metadata.name)
    Mounts:
      /etc/prometheus/config from config (rw)
      /etc/prometheus/config_out from config-out (rw)
      /var/run/secrets/kubernetes.io/serviceaccount from demo-prometheus-operator-prometheus-token-jvbrr (ro)
  rules-configmap-reloader:
    Container ID:  docker://5804e45380ed1b5374a4c2c9ee4c9c4e365bee93b9ccd8b5a21f50886ea81a91
    Image:         quay.io/coreos/configmap-reload:v0.0.1
    Image ID:      docker-pullable://quay.io/coreos/configmap-reload@sha256:e2fd60ff0ae4500a75b80ebaa30e0e7deba9ad107833e8ca53f0047c42c5a057
    Port:          <none>
    Host Port:     <none>
    Args:
      --webhook-url=http://127.0.0.1:9090/-/reload
      --volume-dir=/etc/prometheus/rules/prometheus-demo-prometheus-operator-prometheus-rulefiles-0
    State:          Running
      Started:      Wed, 11 Mar 2020 18:07:06 +0000
    Ready:          True
    Restart Count:  0
    Limits:
      cpu:     100m
      memory:  25Mi
    Requests:
      cpu:        100m
      memory:     25Mi
    Environment:  <none>
    Mounts:
      /etc/prometheus/rules/prometheus-demo-prometheus-operator-prometheus-rulefiles-0 from prometheus-demo-prometheus-operator-prometheus-rulefiles-0 (rw)
      /var/run/secrets/kubernetes.io/serviceaccount from demo-prometheus-operator-prometheus-token-jvbrr (ro)
Conditions:
  Type              Status
  Initialized       True
  Ready             True
  ContainersReady   True
  PodScheduled      True
Volumes:
  config:
    Type:        Secret (a volume populated by a Secret)
    SecretName:  prometheus-demo-prometheus-operator-prometheus
    Optional:    false
  tls-assets:
    Type:        Secret (a volume populated by a Secret)
    SecretName:  prometheus-demo-prometheus-operator-prometheus-tls-assets
    Optional:    false
  config-out:
    Type:       EmptyDir (a temporary directory that shares a pod's lifetime)
    Medium:
    SizeLimit:  <unset>
  prometheus-demo-prometheus-operator-prometheus-rulefiles-0:
    Type:      ConfigMap (a volume populated by a ConfigMap)
    Name:      prometheus-demo-prometheus-operator-prometheus-rulefiles-0
    Optional:  false
  prometheus-demo-prometheus-operator-prometheus-db:
    Type:       EmptyDir (a temporary directory that shares a pod's lifetime)
    Medium:
    SizeLimit:  <unset>
  demo-prometheus-operator-prometheus-token-jvbrr:
    Type:        Secret (a volume populated by a Secret)
    SecretName:  demo-prometheus-operator-prometheus-token-jvbrr
    Optional:    false
QoS Class:       Burstable
Node-Selectors:  <none>
Tolerations:     node.kubernetes.io/not-ready:NoExecute for 300s
                 node.kubernetes.io/unreachable:NoExecute for 300s
Events:
  Type    Reason     Age                    From                                          Message
  ----    ------     ----                   ----                                          -------
  Normal  Scheduled  4m51s                  default-scheduler                             Successfully assigned monitoring/prometheus-demo-prometheus-operator-prometheus-0 to gke-c-7dkls-default-0-c6ca178a-gmcq
  Normal  Pulling    4m45s                  kubelet, gke-c-7dkls-default-0-c6ca178a-gmcq  Pulling image "quay.io/prometheus/prometheus:v2.15.2"
  Normal  Pulled     4m39s                  kubelet, gke-c-7dkls-default-0-c6ca178a-gmcq  Successfully pulled image "quay.io/prometheus/prometheus:v2.15.2"
  Normal  Pulling    4m36s                  kubelet, gke-c-7dkls-default-0-c6ca178a-gmcq  Pulling image "quay.io/coreos/prometheus-config-reloader:v0.37.0"
  Normal  Pulled     4m35s                  kubelet, gke-c-7dkls-default-0-c6ca178a-gmcq  Successfully pulled image "quay.io/coreos/prometheus-config-reloader:v0.37.0"
  Normal  Pulling    4m34s                  kubelet, gke-c-7dkls-default-0-c6ca178a-gmcq  Pulling image "quay.io/coreos/configmap-reload:v0.0.1"
  Normal  Started    4m34s                  kubelet, gke-c-7dkls-default-0-c6ca178a-gmcq  Started container prometheus-config-reloader
  Normal  Created    4m34s                  kubelet, gke-c-7dkls-default-0-c6ca178a-gmcq  Created container prometheus-config-reloader
  Normal  Pulled     4m33s                  kubelet, gke-c-7dkls-default-0-c6ca178a-gmcq  Successfully pulled image "quay.io/coreos/configmap-reload:v0.0.1"
  Normal  Created    4m32s (x2 over 4m36s)  kubelet, gke-c-7dkls-default-0-c6ca178a-gmcq  Created container prometheus
  Normal  Created    4m32s                  kubelet, gke-c-7dkls-default-0-c6ca178a-gmcq  Created container rules-configmap-reloader
  Normal  Started    4m32s                  kubelet, gke-c-7dkls-default-0-c6ca178a-gmcq  Started container rules-configmap-reloader
  Normal  Pulled     4m32s                  kubelet, gke-c-7dkls-default-0-c6ca178a-gmcq  Container image "quay.io/prometheus/prometheus:v2.15.2" already present on machine
  Normal  Started    4m31s (x2 over 4m36s)  kubelet, gke-c-7dkls-default-0-c6ca178a-gmcq  Started container prometheus

让我们清理默认规则，使得我们可以更好地观察我们将要创建的那个规则。以下命令将删除所有规则，但会留下monitoring-demo-prometheus-operator-alertmanager.rules。

$ kubectl -n monitoring delete prometheusrules $(kubectl -n monitoring get prometheusrules | grep -v alert)

$ kubectl -n monitoring get prometheusrules
NAME                                          AGE
demo-prometheus-operator-alertmanager.rules   8m53s

请注意：我们只保留一条规则是为了让demo更容易。但是有一条规则，你绝对不能删除，它位于monitoring-demo-prometheus-operator-general.rules.yaml中，被称为看门狗。该告警总是处于触发状态，其目的是确保整个告警流水线正常运转。

让我们从CLI中检查我们留下的规则并将其与我们将在浏览器中看到的进行比较。

$ kubectl -n monitoring describe prometheusrule demo-prometheus-operator-alertmanager.rules
Name:         demo-prometheus-operator-alertmanager.rules
Namespace:    monitoring
Labels:       app=prometheus-operator
              chart=prometheus-operator-8.12.1
              heritage=Tiller
              release=demo
Annotations:  prometheus-operator-validated: true
API Version:  monitoring.coreos.com/v1
Kind:         PrometheusRule
Metadata:
  Creation Timestamp:  2020-03-11T18:06:25Z
  Generation:          1
  Resource Version:    4871
  Self Link:           /apis/monitoring.coreos.com/v1/namespaces/monitoring/prometheusrules/demo-prometheus-operator-alertmanager.rules
  UID:                 6a84dbb0-feba-4f17-b3dc-4b6486818bc0
Spec:
  Groups:
    Name:  alertmanager.rules
    Rules:
      Alert:  AlertmanagerConfigInconsistent
      Annotations:
        Message:  The configuration of the instances of the Alertmanager cluster `{{$labels.service}}` are out of sync.
      Expr:       count_values("config_hash", alertmanager_config_hash{job="demo-prometheus-operator-alertmanager",namespace="monitoring"}) BY (service) / ON(service) GROUP_LEFT() label_replace(max(prometheus_operator_spec_replicas{job="demo-prometheus-operator-operator",namespace="monitoring",controller="alertmanager"}) by (name, job, namespace, controller), "service", "$1", "name", "(.*)") != 1
      For:        5m
      Labels:
        Severity:  critical
      Alert:       AlertmanagerFailedReload
      Annotations:
        Message:  Reloading Alertmanager's configuration has failed for {{ $labels.namespace }}/{{ $labels.pod}}.
      Expr:       alertmanager_config_last_reload_successful{job="demo-prometheus-operator-alertmanager",namespace="monitoring"} == 0
      For:        10m
      Labels:
        Severity:  warning
      Alert:       AlertmanagerMembersInconsistent
      Annotations:
        Message:  Alertmanager has not found all other members of the cluster.
      Expr:       alertmanager_cluster_members{job="demo-prometheus-operator-alertmanager",namespace="monitoring"}
  != on (service) GROUP_LEFT()
count by (service) (alertmanager_cluster_members{job="demo-prometheus-operator-alertmanager",namespace="monitoring"})
      For:  5m
      Labels:
        Severity:  critical
Events:            <none>

让我们移除所有默认告警并创建一个我们自己的告警：

$ kubectl -n monitoring edit prometheusrules demo-prometheus-operator-alertmanager.rules
prometheusrule.monitoring.coreos.com/demo-prometheus-operator-alertmanager.rules edited

我们的自定义告警如下所示：

$ kubectl -n monitoring describe prometheusrule demo-prometheus-operator-alertmanager.rules
Name:         demo-prometheus-operator-alertmanager.rules
Namespace:    monitoring
Labels:       app=prometheus-operator
              chart=prometheus-operator-8.12.1
              heritage=Tiller
              release=demo
Annotations:  prometheus-operator-validated: true
API Version:  monitoring.coreos.com/v1
Kind:         PrometheusRule
Metadata:
  Creation Timestamp:  2020-03-11T18:06:25Z
  Generation:          3
  Resource Version:    18180
  Self Link:           /apis/monitoring.coreos.com/v1/namespaces/monitoring/prometheusrules/demo-prometheus-operator-alertmanager.rules
  UID:                 6a84dbb0-feba-4f17-b3dc-4b6486818bc0
Spec:
  Groups:
    Name:  alertmanager.rules
    Rules:
      Alert:  PodHighCpuLoad
      Annotations:
        Message:  Alertmanager has found {{ $labels.instance }} with CPU too high
      Expr:       rate (container_cpu_usage_seconds_total{pod_name=~"nginx-.*", image!="", container!="POD"}[5m])  > 0.04
      For:        1m
      Labels:
        Severity:  critical
Events:            <none>

以下是我们创建的告警的选项：

annotation：描述告警的信息标签集。
expr：由PromQL写的表达式
for：可选参数，设置了之后会告诉Prometheus在定义的时间段内告警是否处于active状态。仅在此定义时间后才会触发告警。
label：可以附加到告警的额外标签。如果你想了解更多关于告警的信息，可以访问：

https://prometheus.io/docs/prometheus/latest/configuration/alerting_rules/

现在我们已经完成了Prometheus告警的设置，让我们配置Alertmanager，使得我们能够通过电子邮件获得告警通知。Alertmanager的配置位于Kubernetes secret对象中。

$ kubectl get secrets -n monitoring
NAME                                                        TYPE                                  DATA   AGE
alertmanager-demo-prometheus-operator-alertmanager          Opaque                                1      32m
default-token-x4rgq                                         kubernetes.io/service-account-token   3      37m
demo-grafana                                                Opaque                                3      32m
demo-grafana-test-token-p6qnk                               kubernetes.io/service-account-token   3      32m
demo-grafana-token-ff6nl                                    kubernetes.io/service-account-token   3      32m
demo-kube-state-metrics-token-vmvbr                         kubernetes.io/service-account-token   3      32m
demo-prometheus-node-exporter-token-wlnk9                   kubernetes.io/service-account-token   3      32m
demo-prometheus-operator-admission                          Opaque                                3      32m
demo-prometheus-operator-alertmanager-token-rrx4k           kubernetes.io/service-account-token   3      32m
demo-prometheus-operator-operator-token-q9744               kubernetes.io/service-account-token   3      32m
demo-prometheus-operator-prometheus-token-jvbrr             kubernetes.io/service-account-token   3      32m
prometheus-demo-prometheus-operator-prometheus              Opaque                                1      31m
prometheus-demo-prometheus-operator-prometheus-tls-assets   Opaque                                0      31m

我们只对alertmanager-demo-prometheus-operator-alertmanager感兴趣。让我们看一下：

kubectl -n monitoring get secret alertmanager-demo-prometheus-operator-alertmanager -o yaml
apiVersion: v1
data:
  alertmanager.yaml: Z2xvYmFsOgogIHJlc29sdmVfdGltZW91dDogNW0KcmVjZWl2ZXJzOgotIG5hbWU6ICJudWxsIgpyb3V0ZToKICBncm91cF9ieToKICAtIGpvYgogIGdyb3VwX2ludGVydmFsOiA1bQogIGdyb3VwX3dhaXQ6IDMwcwogIHJlY2VpdmVyOiAibnVsbCIKICByZXBlYXRfaW50ZXJ2YWw6IDEyaAogIHJvdXRlczoKICAtIG1hdGNoOgogICAgICBhbGVydG5hbWU6IFdhdGNoZG9nCiAgICByZWNlaXZlcjogIm51bGwiCg==
kind: Secret
metadata:
  creationTimestamp: "2020-03-11T18:06:24Z"
  labels:
    app: prometheus-operator-alertmanager
    chart: prometheus-operator-8.12.1
    heritage: Tiller
    release: demo
  name: alertmanager-demo-prometheus-operator-alertmanager
  namespace: monitoring
  resourceVersion: "3018"
  selfLink: /api/v1/namespaces/monitoring/secrets/alertmanager-demo-prometheus-operator-alertmanager
  uid: 6baf6883-f690-47a1-bb49-491935956c22
type: Opaque

alertmanager.yaml字段是由base64编码的，让我们看看：

$ echo 'Z2xvYmFsOgogIHJlc29sdmVfdGltZW91dDogNW0KcmVjZWl2ZXJzOgotIG5hbWU6ICJudWxsIgpyb3V0ZToKICBncm91cF9ieToKICAtIGpvYgogIGdyb3VwX2ludGVydmFsOiA1bQogIGdyb3VwX3dhaXQ6IDMwcwogIHJlY2VpdmVyOiAibnVsbCIKICByZXBlYXRfaW50ZXJ2YWw6IDEyaAogIHJvdXRlczoKICAtIG1hdGNoOgogICAgICBhbGVydG5hbWU6IFdhdGNoZG9nCiAgICByZWNlaXZlcjogIm51bGwiCg==' | base64 --decode
global:
  resolve_timeout: 5m
receivers:
- name: "null"
route:
  group_by:
  - job
  group_interval: 5m
  group_wait: 30s
  receiver: "null"
  repeat_interval: 12h
  routes:
  - match:
      alertname: Watchdog
    receiver: "null"

正如我们所看到的，这是默认的Alertmanager配置。你也可以在Alertmanager UI的Status选项卡中查看此配置。接下来，我们来对它进行一些更改——在本例中为发送邮件：

$ cat alertmanager.yaml
global:
  resolve_timeout: 5m
route:
  group_by: [Alertname]
  # Send all notifications to me.
  receiver: demo-alert
  group_wait: 30s
  group_interval: 5m
  repeat_interval: 12h
  routes:
  - match:
      alertname: DemoAlertName
    receiver: 'demo-alert'

receivers:
- name: demo-alert
  email_configs:
  - to: your_email@gmail.com
    from: from_email@gmail.com
    # Your smtp server address
    smarthost: smtp.gmail.com:587
    auth_username: from_email@gmail.com
    auth_identity: from_email@gmail.com
    auth_password: 16letter_generated token # you can use gmail account password, but better create a dedicated token for this
    headers:
      From: from_email@gmail.com
      Subject: 'Demo ALERT'

首先，我们需要对此进行编码：

$  cat alertmanager.yaml | base64 -w0

我们获得编码输出后，我们需要在我们将要应用的yaml文件中填写它：

cat alertmanager-secret-k8s.yaml
apiVersion: v1
data:
  alertmanager.yaml: <paste here de encoded content of alertmanager.yaml>
kind: Secret
metadata:
  name: alertmanager-demo-prometheus-operator-alertmanager
  namespace: monitoring
type: Opaque

$ kubectl apply -f alertmanager-secret-k8s.yaml
Warning: kubectl apply should be used on resource created by either kubectl create --save-config or kubectl apply
secret/alertmanager-demo-prometheus-operator-alertmanager configured

该配置将会自动重新加载并在UI中显示更改。

接下来，我们部署一些东西来对其进行监控。对于本例而言，一个简单的nginx deployment已经足够：

$ cat nginx-deployment.yaml
apiVersion: apps/v1 # for versions before 1.9.0 use apps/v1beta2
kind: Deployment
metadata:
  name: nginx-deployment
spec:
  selector:
    matchLabels:
      app: nginx
  replicas: 3 # tells deployment to run 2 pods matching the template
  template:
    metadata:
      labels:
        app: nginx
    spec:
      containers:
      - name: nginx
        image: nginx:1.7.9
        ports:
        - containerPort: 80

$ kubectl apply -f nginx-deployment.yaml
deployment.apps/nginx-deployment created

根据配置yaml，我们有3个副本：

$ kubectl get pods
NAME                                READY   STATUS    RESTARTS   AGE
nginx-deployment-5754944d6c-7g6gq   1/1     Running   0          67s
nginx-deployment-5754944d6c-lhvx8   1/1     Running   0          67s
nginx-deployment-5754944d6c-whhtr   1/1     Running   0          67s

在Prometheus UI中，使用我们为告警配置的相同表达式：

rate (container_cpu_usage_seconds_total{pod_name=~"nginx-.*", image!="", container!="POD"}[5m])

我们可以为这些Pod检查数据，所有Pod的值应该为0。

让我们在其中一个pod中添加一些负载，然后来看看值的变化，当值大于0.04时，我们应该接收到告警：

$  kubectl exec -it nginx-deployment-5754944d6c-7g6gq -- /bin/sh
# yes > /dev/null

该告警有3个阶段：

Inactive：不满足告警触发条件
Pending：条件已满足
Firing：触发告警

我们已经看到告警处于inactive状态，所以在CPU上添加一些负载，以观察到剩余两种状态：

告警一旦触发，将会在Alertmanager中显示：

Alertmanger配置为当我们收到告警时发送邮件。所以此时，如果我们检查收件箱，会看到类似以下内容：

总结

我们知道监控的重要性，但是如果没有告警，它将是不完整的。发生问题时，告警可以立即通知我们，让我们立即了解到系统出现了问题。而Prometheus涵盖了这两个方面：既有监控解决方案又通过Alertmanager组件发出告警。本文中，我们看到了如何在Prometheus配置中定义告警以及告警在触发时如何到达Alertmanager。然后根据Alertmanager的定义/集成，我们收到了一封电子邮件，其中包含触发的告警的详细信息（也可以通过Slack或PagerDuty发送）。

创作场景

如何使用 Prometheus 配置自定义告警规则