kubernetes - 如何配置Helm在kubernetes上安装的alertmanager？

Question

在kubernetes集群中使用已安装的Helm和Prometheus的Grafana:

helm install stable/prometheus
helm install stable/grafana

它具有 alertmanage服务。



但是我看到一个博客介绍了如何使用yaml文件设置alertmanager config:

http://blog.wercker.com/how-to-setup-alerts-on-prometheus

是否可以使用当前方式(由helm安装)来为 alert rules， CPU设置一些 memory和配置并发送电子邮件，而无需创建其他yaml文件？

我看到了k8s configmap到 alertmanager的介绍:

https://github.com/kubernetes/charts/tree/master/stable/prometheus#configmap-files

但不清楚如何使用和怎么做。

编辑

我下载了 stable/prometheus的源代码以查看其作用。从 values.yaml文件中，我发现:

serverFiles:
  alerts: ""
  rules: ""

  prometheus.yml: |-
    rule_files:
      - /etc/config/rules
      - /etc/config/alerts
    scrape_configs:
      - job_name: prometheus
        static_configs:
          - targets:
            - localhost:9090

https://github.com/kubernetes/charts/blob/master/stable/prometheus/values.yaml#L600

因此，我认为应该自己写入此配置文件以在此处定义警报 rules和 alertmanager。但是不要清除此块:

    rule_files:
      - /etc/config/rules
      - /etc/config/alerts

也许这意味着容器中的路径。但是现在没有任何文件。应该在这里添加:

serverFiles:
  alert: ""
  rules: ""

编辑2

在 alert rules中设置 alertmanager和 values.yaml配置后:

## Prometheus server ConfigMap entries
##
serverFiles:
  alerts: ""
  rules: |-
    #
    # CPU Alerts
    #
    ALERT HighCPU
      IF ((sum(node_cpu{mode=~"user|nice|system|irq|softirq|steal|idle|iowait"}) by (instance, job)) - ( sum(node_cpu{mode=~"idle|iowait"}) by (instance,job) )   )   /  (sum(node_cpu{mode=~"user|nice|system|irq|softirq|steal|idle|iowait"}) by (instance, job)) * 100 > 95
      FOR 10m
      LABELS { service = "backend" }
      ANNOTATIONS {
        summary = "High CPU Usage",
        description = "This machine  has really high CPU usage for over 10m",
      }

    # TEST
    ALERT APIHighRequestLatency
      IF api_http_request_latencies_second{quantile="0.5"} >1
      FOR 1m
      ANNOTATIONS {
        summary = "High request latency on {{$labels.instance }}",
        description = "{{ $labels.instance }} has amedian request latency above 1s (current value: {{ $value }}s)",
    }

运行 helm install prometheus/进行安装。

启动 port-forward组件的 alertmanager:

export POD_NAME=$(kubectl get pods --namespace default -l "app=prometheus,component=alertmanager" -o jsonpath="{.items[0].metadata.name}")
kubectl --namespace default port-forward $POD_NAME 9093

然后访问浏览器 http://127.0.0.1:9003，得到以下消息:

Forwarding from 127.0.0.1:9093 -> 9093
Handling connection for 9093
Handling connection for 9093
E0122 17:41:53.229084    7159 portforward.go:331] an error occurred forwarding 9093 -> 9093: error forwarding port 9093 to pod 6614ee96df545c266e5fff18023f8f7c87981f3340ee8913acf3d8da0e39e906, uid : exit status 1: 2018/01/22 08:37:54 socat[31237.140275133073152] E connect(5, AF=2 127.0.0.1:9093, 16): Connection refused
Handling connection for 9093
E0122 17:41:53.243511    7159 portforward.go:331] an error occurred forwarding 9093 -> 9093: error forwarding port 9093 to pod 6614ee96df545c266e5fff18023f8f7c87981f3340ee8913acf3d8da0e39e906, uid : exit status 1: 2018/01/22 08:37:54 socat[31238.140565602109184] E connect(5, AF=2 127.0.0.1:9093, 16): Connection refused
E0122 17:41:53.246011    7159 portforward.go:331] an error occurred forwarding 9093 -> 9093: error forwarding port 9093 to pod 6614ee96df545c266e5fff18023f8f7c87981f3340ee8913acf3d8da0e39e906, uid : exit status 1: 2018/01/22 08:37:54 socat[31239.140184300869376] E connect(5, AF=2 127.0.0.1:9093, 16): Connection refused
Handling connection for 9093
Handling connection for 9093
E0122 17:41:53.846399    7159 portforward.go:331] an error occurred forwarding 9093 -> 9093: error forwarding port 9093 to pod 6614ee96df545c266e5fff18023f8f7c87981f3340ee8913acf3d8da0e39e906, uid : exit status 1: 2018/01/22 08:37:55 socat[31250.140004515874560] E connect(5, AF=2 127.0.0.1:9093, 16): Connection refused
E0122 17:41:53.847821    7159 portforward.go:331] an error occurred forwarding 9093 -> 9093: error forwarding port 9093 to pod 6614ee96df545c266e5fff18023f8f7c87981f3340ee8913acf3d8da0e39e906, uid : exit status 1: 2018/01/22 08:37:55 socat[31251.140355466835712] E connect(5, AF=2 127.0.0.1:9093, 16): Connection refused
Handling connection for 9093
E0122 17:41:53.858521    7159 portforward.go:331] an error occurred forwarding 9093 -> 9093: error forwarding port 9093 to pod 6614ee96df545c266e5fff18023f8f7c87981f3340ee8913acf3d8da0e39e906, uid : exit status 1: 2018/01/22 08:37:55 socat[31252.140268300003072] E connect(5, AF=2 127.0.0.1:9093, 16): Connection refused

为什么？

当我检查 kubectl describe po illocutionary-heron-prometheus-alertmanager-587d747b9c-qwmm6时，得到了:

Name:           illocutionary-heron-prometheus-alertmanager-587d747b9c-qwmm6
Namespace:      default
Node:           minikube/192.168.99.100
Start Time:     Mon, 22 Jan 2018 17:33:54 +0900
Labels:         app=prometheus
                component=alertmanager
                pod-template-hash=1438303657
                release=illocutionary-heron
Annotations:    kubernetes.io/created-by={"kind":"SerializedReference","apiVersion":"v1","reference":{"kind":"ReplicaSet","namespace":"default","name":"illocutionary-heron-prometheus-alertmanager-587d747b9c","uid":"f...
Status:         Running
IP:             172.17.0.10
Created By:     ReplicaSet/illocutionary-heron-prometheus-alertmanager-587d747b9c
Controlled By:  ReplicaSet/illocutionary-heron-prometheus-alertmanager-587d747b9c
Containers:
  prometheus-alertmanager:
    Container ID:  docker://0808a3ecdf1fa94b36a1bf4b8f0d9d2933bc38afa8b25e09d0d86f036ac3165b
    Image:         prom/alertmanager:v0.9.1
    Image ID:      docker-pullable://prom/alertmanager@sha256:ed926b227327eecfa61a9703702c9b16fc7fe95b69e22baa656d93cfbe098320
    Port:          9093/TCP
    Args:
      --config.file=/etc/config/alertmanager.yml
      --storage.path=/data
    State:          Waiting
      Reason:       CrashLoopBackOff
    Last State:     Terminated
      Reason:       Error
      Exit Code:    1
      Started:      Mon, 22 Jan 2018 17:55:24 +0900
      Finished:     Mon, 22 Jan 2018 17:55:24 +0900
    Ready:          False
    Restart Count:  9
    Readiness:      http-get http://:9093/%23/status delay=30s timeout=30s period=10s #success=1 #failure=3
    Environment:    <none>
    Mounts:
      /data from storage-volume (rw)
      /etc/config from config-volume (rw)
      /var/run/secrets/kubernetes.io/serviceaccount from default-token-h5b8l (ro)
  prometheus-alertmanager-configmap-reload:
    Container ID:  docker://b4a349bf7be4ea78abe6899ad0173147f0d3f6ff1005bc513b2c0ac726385f0b
    Image:         jimmidyson/configmap-reload:v0.1
    Image ID:      docker-pullable://jimmidyson/configmap-reload@sha256:2d40c2eaa6f435b2511d0cfc5f6c0a681eeb2eaa455a5d5ac25f88ce5139986e
    Port:          <none>
    Args:
      --volume-dir=/etc/config
      --webhook-url=http://localhost:9093/-/reload
    State:          Running
      Started:      Mon, 22 Jan 2018 17:33:56 +0900
    Ready:          True
    Restart Count:  0
    Environment:    <none>
    Mounts:
      /etc/config from config-volume (ro)
      /var/run/secrets/kubernetes.io/serviceaccount from default-token-h5b8l (ro)
Conditions:
  Type           Status
  Initialized    True 
  Ready          False 
  PodScheduled   True 
Volumes:
  config-volume:
    Type:      ConfigMap (a volume populated by a ConfigMap)
    Name:      illocutionary-heron-prometheus-alertmanager
    Optional:  false
  storage-volume:
    Type:       PersistentVolumeClaim (a reference to a PersistentVolumeClaim in the same namespace)
    ClaimName:  illocutionary-heron-prometheus-alertmanager
    ReadOnly:   false
  default-token-h5b8l:
    Type:        Secret (a volume populated by a Secret)
    SecretName:  default-token-h5b8l
    Optional:    false
QoS Class:       BestEffort
Node-Selectors:  <none>
Tolerations:     <none>
Events:
  Type     Reason                 Age                 From               Message
  ----     ------                 ----                ----               -------
  Warning  FailedScheduling       29m (x2 over 29m)   default-scheduler  PersistentVolumeClaim is not bound: "illocutionary-heron-prometheus-alertmanager"
  Normal   Scheduled              29m                 default-scheduler  Successfully assigned illocutionary-heron-prometheus-alertmanager-587d747b9c-qwmm6 to minikube
  Normal   SuccessfulMountVolume  29m                 kubelet, minikube  MountVolume.SetUp succeeded for volume "config-volume"
  Normal   SuccessfulMountVolume  29m                 kubelet, minikube  MountVolume.SetUp succeeded for volume "pvc-fa84b197-ff4e-11e7-a584-0800270fb7fc"
  Normal   SuccessfulMountVolume  29m                 kubelet, minikube  MountVolume.SetUp succeeded for volume "default-token-h5b8l"
  Normal   Started                29m                 kubelet, minikube  Started container
  Normal   Created                29m                 kubelet, minikube  Created container
  Normal   Pulled                 29m                 kubelet, minikube  Container image "jimmidyson/configmap-reload:v0.1" already present on machine
  Normal   Started                29m (x3 over 29m)   kubelet, minikube  Started container
  Normal   Created                29m (x4 over 29m)   kubelet, minikube  Created container
  Normal   Pulled                 29m (x4 over 29m)   kubelet, minikube  Container image "prom/alertmanager:v0.9.1" already present on machine
  Warning  BackOff                9m (x91 over 29m)   kubelet, minikube  Back-off restarting failed container
  Warning  FailedSync             4m (x113 over 29m)  kubelet, minikube  Error syncing pod

编辑3
alertmanager文件中的 values.yaml配置:

## alertmanager ConfigMap entries
##
alertmanagerFiles:
  alertmanager.yml: |-
    global:
      resolve_timeout: 5m
      smtp_smarthost: smtp.gmail.com:587
      smtp_from: sender@gmail.com
      smtp_auth_username: sender@gmail.com
      smtp_auth_password: sender_password

    receivers:
      - name: default-receiver
      email_configs:
      - to: target_email@gmail.com

    route:
      group_wait: 10s
      group_interval: 5m
      receiver: default-receiver
      repeat_interval: 3h

不行。上面有错误。

alertmanagerFiles:
  alertmanager.yml: |-
    global:
      # slack_api_url: ''

    receivers:
      - name: default-receiver
        # slack_configs:
        #  - channel: '@you'
        #    send_resolved: true

    route:
      group_wait: 10s
      group_interval: 5m
      receiver: default-receiver
      repeat_interval

正常运行，没有任何错误。

因此，问题出在 email_configs配置方法。

Answer 1

alerts文件的rules组中的serverFiles和values.yaml键安装在/etc/config文件夹中的Prometheus容器中。您可以在其中放置所需的配置(例如，从链接的博客文章中汲取灵感)，Prometheus将使用它来处理警报。

例如，可以这样设置一个简单的规则:

serverFiles:
  alerts: |
    ALERT cpu_threshold_exceeded   
    IF (100 * (1 - avg by(job)(irate(node_cpu{mode='idle'}[5m])))) > 80
    FOR 300s
    LABELS {
      severity = "warning",
    }
    ANNOTATIONS {
      summary = "CPU usage > 80% for {{ $labels.job }}",
      description = "CPU usage avg for last 5m: {{ $value }}",
    }