实用干货丨如何使用Prometheus配置自定义告警规则

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：

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）。