一、监控k8s集群
要监控k8s集群需要使用到以下服务用于收集监控的资源信息，node_exporter用于监控k8s集群节点的资源信息，kube-state-metrics用于监控k8s集群的deployment、statefulset、daemonset、pod等的状态，cadvisor用于监控k8s集群的pod资源信息
在k8s集群中创建monitoring的命名空间用于部署监控的容器
kubectl create namespace monitoring
在k8s集群中部署node_exporter容器服务
vinode-exporter.yaml
apiVersion: apps/v1kind: DaemonSet#使用daemonset控制器，使得集群中的每个节点都能部署一个podmetadata:name: node-exporternamespace: monitoring labels:k8s-app: node-exporterspec:selector:matchLabels:k8s-app: node-exportertemplate:metadata:labels:k8s-app: node-exporterspec:tolerations:#配置容忍策略，使得pod能部署在master节点上- effect: NoSchedulekey: node-role.kubernetes.io/control-planecontainers:- image: prom/node-exporter#配置node-exporter的镜像imagePullPolicy: IfNotPresentname: prometheus-node-exporterports:- containerPort: 9100 #配置容器端口hostPort: 9100 #配置绑定k8s主机节点的端口，用于提供对外访问的接口protocol: TCPname: metricshostNetwork: true#使用hostNetwork: true是必要的，这样才能将Pod的网络栈绑定到宿主机上，以实现hostPort的功能执行yaml生成node-exporter容器
kubectl apply -ynode-exporter.yaml
查看容器
kubectl get pod -n monitoring -lk8s-app=node-exporter -o wide
可以看到集群的每个节点都有一个node_exporter的pod服务
查看收集的数据
http://10.1.60.119:9100/metrics
在k8s集群中部署kube-state-metrics容器服务
部署kube-state-metrics服务需要去github上的项目拉取yaml
下载地址：https://github.com/kubernetes/kube-state-metrics/tree/v2.9.2
需要根据自己的k8s集群版本下载合适的kube-state-metrics版本，我的k8s版本是1.26.0所以我是下载了2.9.2版本的kube-state-metrics
mkdir /opt/kube-state-metrics && cd/opt/kube-state-metrics
将下载的安装包放到该目录下解压
tar -zxvfkube-state-metrics-2.9.2.tar.gz
将需要用到的yaml文件拷贝出来
mvkube-state-metrics-2.9.2/examples/standard/* /opt/kube-state-metrics
ls
更改一下yaml文件
vi deployment.yaml
apiVersion: apps/v1kind: Deployment#使用deployment控制器，将pod部署在工作节点即可metadata:labels:app.kubernetes.io/component: exporterapp.kubernetes.io/name: kube-state-metricsapp.kubernetes.io/version: 2.9.2name: kube-state-metricsnamespace: kube-systemspec:replicas: 1selector:matchLabels:app.kubernetes.io/name: kube-state-metricstemplate:metadata:labels:app.kubernetes.io/component: exporterapp.kubernetes.io/name: kube-state-metricsapp.kubernetes.io/version: 2.9.2spec:automountServiceAccountToken: truecontainers:- image: bitnami/kube-state-metrics:2.9.2 #更改镜像地址，原本的镜像在国外拉不下来livenessProbe:httpGet:path: /healthzport: 8080initialDelaySeconds: 5timeoutSeconds: 5name: kube-state-metricsports:- containerPort: 8080name: http-metrics- containerPort: 8081name: telemetryreadinessProbe:httpGet:path: /port: 8081initialDelaySeconds: 5timeoutSeconds: 5securityContext:allowPrivilegeEscalation: falsecapabilities:drop:- ALLreadOnlyRootFilesystem: truerunAsNonRoot: truerunAsUser: 65534seccompProfile:type: RuntimeDefaultnodeSelector:kubernetes.io/os: linuxserviceAccountName: kube-state-metrics关于镜像的问题可以使用docker命令查一下镜像
docker searchkube-state-metrics
viservice.yaml
apiVersion: v1kind: Servicemetadata:labels:app.kubernetes.io/component: exporterapp.kubernetes.io/name: kube-state-metricsapp.kubernetes.io/version: 2.9.2name: kube-state-metricsnamespace: kube-systemspec:type: NodePortclusterIP:ports:- name: http-metricsport: 8080nodePort: 30080#原本的端口值比较大，超过了nodeport的端口范围targetPort: http-metricsprotocol: TCP- name: telemetryport: 8081 nodePort: 30081#原本的端口值比较大，超过了nodeport的端口范围targetPort: telemetryprotocol: TCPselector:app.kubernetes.io/name: kube-state-metrics其它的yaml保持默认即可
执行yaml创建kube-state-metrics服务
kubectl apply -f/opt/kube-state-metrics/
查看pod、svc服务
kubectl get pod,svc-n kube-system
查看收集的数据
http://10.1.60.119:30080/metrics
在k8s集群中部署cadvisor容器服务
vicadvisor.yaml
apiVersion: apps/v1kind: DaemonSetmetadata:name: cadvisornamespace: monitoringspec:selector:matchLabels:k8s-app: cadvisortemplate:metadata:labels:k8s-app: cadvisorspec:tolerations:- effect: NoSchedulekey: node-role.kubernetes.io/control-planehostNetwork: truerestartPolicy: Alwayscontainers:- name: cadvisorimage: google/cadvisorimagePullPolicy: IfNotPresentports:- containerPort: 8080hostPort: 8080protocol: TCPname: metrics执行yaml生成cadvisor容器
kubectl apply -ycadvisor.yaml
查看容器
kubectl get pod -n monitoring -lk8s-app=cadvisor -o wide
可以看到集群的每个节点都有一个cadvisor的pod服务
查看收集的数据
http://10.1.60.119:8080/metrics
二、Prometheus获取监控服务的数据并使用grafana展示
部署prometheus
参考：prometheus部署_Apex Predator的博客-CSDN博客
部署grafana
参考：grafana部署_Apex Predator的博客-CSDN博客
配置prometheus
编辑Prometheus配置文件
vi /opt/prometheus/prometheus/prometheus.yml
global:scrape_interval: 15s evaluation_interval: 15s #alerting:#关于告警组件的配置先忽略 # alertmanagers: # - static_configs:#- targets:#- 10.1.60.118:9093#rule_files: #关于告警规则的配置先忽略#- "/opt/prometheus/prometheus/rule/*.yml"scrape_configs:- job_name: "prometheus"static_configs:- targets: ["localhost:9090"]- job_name: "k8s_node_exporter"#配置k8s集群node_exporter监控数据服务的接口static_configs:- targets: ["10.1.60.119:9100","10.1.60.120:9100","10.1.60.121:9100","10.1.60.122:9100","10.1.60.123:9100"]- job_name: "k8s_pod_cadvisor" #配置k8s集群cadvisor监控数据服务的接口static_configs:- targets: ["10.1.60.119:8080","10.1.60.120:8080","10.1.60.121:8080","10.1.60.122:8080","10.1.60.123:8080"]- job_name: "kube-state-metrics" #配置k8s集群kube-state-metrics监控数据服务的接口static_configs:- targets: ["10.1.60.119:30081"]- job_name: "kube-state-telemetry"static_configs:- targets: ["10.1.60.119:30080"]重启prometheus服务
systemctl restart prometheus
查看prometheus监控接口的情况
http://10.1.60.118:9090
配置grafana
配置prometheus为数据源
配置数据展示的dashboard
在以下网页中找到需要的模板
地址：Dashboards | Grafana Labs
node_exporter服务的模板我们就使用id为1860的模板
kube-state-metrics服务的模板我们就使用id为13332的模板
cadvisor服务的模板我们就使用id为1860的模板
配置grafana应用模板
其余两个也是一样找到模板id后进行配置即可，这里就不再展示了
三、Prometheus配置告警规则和告警服务实现钉钉告警
要实现钉钉告警需要部署alertmanager和prometheus-webhook-dingtalk服务
部署参考：prometheus告警发送组件部署_Apex Predator的博客-CSDN博客
配置prometheus告警规则
关于prometheus的告警规则可以在以下网站中找，里面有很多的告警规则
参考：Awesome Prometheus alerts | Collection of alerting rules
我这里就配置k8s集群主机节点的告警规则和pod的一些告警规则
mkdir /opt/prometheus/prometheus/rule && cd/opt/prometheus/prometheus/rule
vi node_exporter.yml
groups:- name: 服务器资源监控rules:- alert: 内存使用率过高expr: 100 - (node_memory_MemAvailable_bytes / node_memory_MemTotal_bytes) * 100 > 80for: 3mlabels:severity: 严重告警annotations:summary: "{{ $labels.instance }} 内存使用率过高, 请尽快处理！"description: "{{ $labels.instance }}内存使用率超过80%,当前使用率{{ $value }}%."- alert: 服务器宕机expr: up == 0for: 1slabels:severity: 严重告警annotations:summary: "{{$labels.instance}} 服务器宕机, 请尽快处理!"description: "{{$labels.instance}} 服务器延时超过3分钟,当前状态{{ $value }}. "- alert: CPU高负荷expr: 100 - (avg by (instance,job)(irate(node_cpu_seconds_total{mode="idle"}[5m])) * 100) > 90for: 5mlabels:severity: 严重告警annotations:summary: "{{$labels.instance}} CPU使用率过高,请尽快处理！"description: "{{$labels.instance}} CPU使用大于90%,当前使用率{{ $value }}%. "- alert: 磁盘IO性能expr: avg(irate(node_disk_io_time_seconds_total[1m])) by(instance,job)* 100 > 90for: 5mlabels:severity: 严重告警annotations:summary: "{{$labels.instance}} 流入磁盘IO使用率过高,请尽快处理！"description: "{{$labels.instance}} 流入磁盘IO大于90%,当前使用率{{ $value }}%."- alert: 网络流入expr: ((sum(rate (node_network_receive_bytes_total{device!~'tap.*|veth.*|br.*|docker.*|virbr*|lo*'}[5m])) by (instance,job)) / 100) > 102400for: 5mlabels:severity: 严重告警annotations:summary: "{{$labels.instance}} 流入网络带宽过高，请尽快处理！"description: "{{$labels.instance}} 流入网络带宽持续5分钟高于100M. RX带宽使用量{{$value}}." - alert: 网络流出expr: ((sum(rate (node_network_transmit_bytes_total{device!~'tap.*|veth.*|br.*|docker.*|virbr*|lo*'}[5m])) by (instance,job)) / 100) > 102400for: 5mlabels:severity: 严重告警annotations:summary: "{{$labels.instance}} 流出网络带宽过高,请尽快处理！"description: "{{$labels.instance}} 流出网络带宽持续5分钟高于100M. RX带宽使用量{$value}}."- alert: TCP连接数expr: node_netstat_Tcp_CurrEstab > 10000for: 2mlabels:severity: 严重告警annotations:summary: " TCP_ESTABLISHED过高！"description: "{{$labels.instance}} TCP_ESTABLISHED大于100%,当前使用率{{ $value }}%." - alert: 磁盘容量expr: 100-(node_filesystem_free_bytes{fstype=~"ext4|xfs"}/node_filesystem_size_bytes {fstype=~"ext4|xfs"}*100) > 90for: 1mlabels:severity: 严重告警annotations:summary: "{{$labels.mountpoint}} 磁盘分区使用率过高，请尽快处理！"description: "{{$labels.instance}} 磁盘分区使用大于90%，当前使用率{{ $value }}%."vikube-state-metrics.yml
groups: #用于定义一个或多个告警规则分组- name: k8s容器服务监控#告警规则分组的名称，用于标识一组相关的告警规则rules:#规则列表，每个规则定义了一个具体的告警条件和处理方式- alert: KubernetesNodeNotReady#告警规则的名称，用于标识告警规则expr: kube_node_status_condition{condition="Ready",status="true"} == 0#定义告警的条件for: 10m#告警规则的持续时间配置，规定了节点状态满足告警条件的持续时间达到 10 分钟时触发告警labels:severity: 严重告警 #告警规则标签annotations:summary: "Kubernetes node not ready (instance {{ $labels.instance }})"description: "Node {{ $labels.node }} has been unready for a long time\nVALUE = {{ $value }}\nLABELS = {{ $labels }}"- alert: KubernetesOutOfCapacityexpr: sum by (node) ((kube_pod_status_phase{phase="Running"} == 1) + on(uid) group_left(node) (0 * kube_pod_info{pod_template_hash=""})) / sum by (node) (kube_node_status_allocatable{resource="pods"}) * 100 > 90for: 2mlabels:severity: 严重告警annotations:summary: "Kubernetes out of capacity (instance {{ $labels.instance }})"description: "{{ $labels.node }} is out of capacity\nVALUE = {{ $value }}\nLABELS = {{ $labels }}"- alert: KubernetesContainerOomKillerexpr: (kube_pod_container_status_restarts_total - kube_pod_container_status_restarts_total offset 10m >= 1) and ignoring (reason) min_over_time(kube_pod_container_status_last_terminated_reason{reason="OOMKilled"}[10m]) == 1for: 0mlabels:severity: 严重告警annotations:summary: "Kubernetes container oom killer (instance {{ $labels.instance }})"description: "Container {{ $labels.container }} in pod {{ $labels.namespace }}/{{ $labels.pod }} has been OOMKilled {{ $value }} times in the last 10 minutes.\nVALUE = {{ $value }}\nLABELS = {{ $labels }}"- alert: KubernetesVolumeOutOfDiskSpaceexpr: kubelet_volume_stats_available_bytes / kubelet_volume_stats_capacity_bytes * 100 < 10for: 2mlabels:severity: 严重告警annotations:summary: "Kubernetes Volume out of disk space (instance {{ $labels.instance }})"description: "Volume is almost full (< 10% left)\nVALUE = {{ $value }}\nLABELS = {{ $labels }}"- alert: KubernetesPersistentvolumeErrorexpr: kube_persistentvolume_status_phase{phase=~"Failed|Pending", job="kube-state-metrics"} > 0for: 0mlabels:severity: 严重告警annotations:summary: "Kubernetes PersistentVolume error (instance {{ $labels.instance }})"description: "Persistent volume is in bad state\nVALUE = {{ $value }}\nLABELS = {{ $labels }}"- alert: pod不健康expr: sum by (namespace, pod) (kube_pod_status_phase{phase=~"Pending|Unknown|Failed"}) > 0for: 15mlabels:severity: 严重告警annotations:summary: "Kubernetes Pod not healthy (instance {{ $labels.instance }})"description: "Pod has been in a non-ready state for longer than 15 minutes.\nVALUE = {{ $value }}\nLABELS = {{ $labels }}"- alert: pod循环重启expr: increase(kube_pod_container_status_restarts_total[2m]) > 1for: 0mlabels:severity: 严重告警annotations:summary: "Kubernetes pod crash looping (instance {{ $labels.instance }})"description: "Pod {{ $labels.pod }} is crash looping\nVALUE = {{ $value }}\nLABELS = {{ $labels }}"- alert: deployment部署失败未回滚expr: kube_deployment_status_observed_generation != kube_deployment_metadata_generationfor: 10mlabels:severity: 严重告警annotations:summary: "Kubernetes Deployment generation mismatch (instance {{ $labels.instance }})"description: "A Deployment has failed but has not been rolled back.\nVALUE = {{ $value }}\nLABELS = {{ $labels }}"- alert: k8s证书临期警告expr: apiserver_client_certificate_expiration_seconds_count{job="apiserver"} > 0 and histogram_quantile(0.01, sum by (job, le) (rate(apiserver_client_certificate_expiration_seconds_bucket{job="apiserver"}[5m]))) < 7*24*60*60for: 0mlabels:severity: 严重告警annotations:summary: "Kubernetes client certificate expires next week (instance {{ $labels.instance }})"description: "A client certificate used to authenticate to the apiserver is expiring next week.\nVALUE = {{ $value }}\nLABELS = {{ $labels }}"重启Prometheus服务
systemctl restart prometheus
访问prometheus查看规则是否生效
http://10.1.60.118:9090
告警测试，关闭node_exporter服务，看看是否会告警
kubectl delete -f node_exporter.yaml
通过prometheus我们可以看到告警规则首先是变成了pending状态
然后等了一会后转变为firing状态，这是因为配置了for，当触发条件满足一段时间后才会完全转化为触发告警
等待30s后将会收到钉钉告警，这是因为alertmanager配置了group_wait，当一组告警被触发后，在这个时间段内，其他属于同一组的告警也会被等待。这可以用于在一定时间内收集同一组告警，以便一次性发送通知
现在来将服务恢复一下，看多久会告警
kubectl apply -fnode_exporter.yaml
可以看到是间隔时间几分钟后才告警恢复，这是因为alertmanager配置了group_interval，一旦一个告警组的首个告警触发了通知，等待指定的间隔时间后，即使组内有其他告警，也会重新触发通知。这可以避免过于频繁地发送通知
其他的告警规则服务我就不一个一个测试了，都是没有问题的