kubernetes-x/kubernetes-MD/基于kubernetes部署Prometheus和Gr...

684 lines
26 KiB
Markdown
Raw Normal View History

2023-03-25 22:18:54 +08:00
<h1><center>基于kubernetes部署Prometheus和Grafana</center></h1>
著作:行癫 <盗版必究>
------
## 一:环境准备
#### 1.kubernetes集群正常
```shell
[root@master ~]# kubectl get node
NAME STATUS ROLES AGE VERSION
master Ready control-plane,master 36d v1.23.1
node-1 Ready <none> 36d v1.23.1
node-2 Ready <none> 36d v1.23.1
node-3 Ready <none> 36d v1.23.1
```
#### 2.harbor仓库正常
![image-20220602010601512](%E5%9F%BA%E4%BA%8Ekubernetes%E9%83%A8%E7%BD%B2Prometheus%E5%92%8CGrafana.assets/image-20220602010601512.png)
## 二Prometheus部署
#### 1.node-exporter部署
node-exporter可以采集机器物理机、虚拟机、云主机等的监控指标数据能够采集到的指标包括CPU, 内存,磁盘,网络,文件数等信息
创建监控namespace
```shell
[root@master ~]# kubectl create ns monitor-sa
```
创建node-export.yaml
```shell
[root@master ~]# vim node-export.yaml
apiVersion: apps/v1
kind: DaemonSet # 可以保证k8s集群的每个节点都运行完全一样的pod
metadata:
name: node-exporter
namespace: monitor-sa
labels:
name: node-exporter
spec:
selector:
matchLabels:
name: node-exporter
template:
metadata:
labels:
name: node-exporter
spec:
hostPID: true
hostIPC: true
hostNetwork: true
containers:
- name: node-exporter
image: prom/node-exporter:v0.16.0
#image: 10.0.0.230/xingdian/node-exporter:v0.16.0
ports:
- containerPort: 9100
resources:
requests:
cpu: 0.15 # 这个容器运行至少需要0.15核cpu
securityContext:
privileged: true # 开启特权模式
args:
- --path.procfs
- /host/proc
- --path.sysfs
- /host/sys
- --collector.filesystem.ignored-mount-points
- '"^/(sys|proc|dev|host|etc)($|/)"'
volumeMounts:
- name: dev
mountPath: /host/dev
- name: proc
mountPath: /host/proc
- name: sys
mountPath: /host/sys
- name: rootfs
mountPath: /rootfs
tolerations:
- key: "node-role.kubernetes.io/master"
operator: "Exists"
effect: "NoSchedule"
volumes:
- name: proc
hostPath:
path: /proc
- name: dev
hostPath:
path: /dev
- name: sys
hostPath:
path: /sys
- name: rootfs
hostPath:
path: /
```
注意:
hostNetwork、hostIPC、hostPID都为True时表示这个Pod里的所有容器会直接使用宿主机的网络直接与宿主机进行IPC进程间通信通信可以看到宿主机里正在运行的所有进程。加入了hostNetwork:true会直接将我们的宿主机的9100端口映射出来从而不需要创建service 在我们的宿主机上就会有一个9100的端口
创建:
```shell
[root@master ~]# kubectl apply -f node-export.yaml
```
查看node-exporter是否部署成功
```shell
[root@master ~]# kubectl get pods -n monitor-sa -o wide
NAME READY STATUS RESTARTS AGE IP NODE NOMINATED NODE READINESS GATES
node-exporter-2cbrg 1/1 Running 0 34m 10.0.0.220 master <none> <none>
node-exporter-7rrbh 1/1 Running 0 34m 10.0.0.222 node-2 <none> <none>
node-exporter-96v29 1/1 Running 0 34m 10.0.0.221 node-1 <none> <none>
node-exporter-bf2j8 1/1 Running 0 34m 10.0.0.223 node-3 <none> <none>
```
注意:
node-export默认的监听端口是9100可以看到当前主机获取到的所有监控数据
```shell
[root@master ~]# curl http://10.0.0.220:9100/metrics | grep node_cpu_seconds
% Total % Received % Xferd Average Speed Time Time Time Current
Dload Upload Total Spent Left Speed
0 0 0 0 0 0 0 0 --:--:-- --:--:-- --:--:-- 0# HELP node_cpu_seconds_total Seconds the cpus spent in each mode.
# TYPE node_cpu_seconds_total counter
node_cpu_seconds_total{cpu="0",mode="idle"} 8398.49
node_cpu_seconds_total{cpu="0",mode="iowait"} 1.54
node_cpu_seconds_total{cpu="0",mode="irq"} 0
node_cpu_seconds_total{cpu="0",mode="nice"} 0
node_cpu_seconds_total{cpu="0",mode="softirq"} 17.2
node_cpu_seconds_total{cpu="0",mode="steal"} 0
node_cpu_seconds_total{cpu="0",mode="system"} 70.61
node_cpu_seconds_total{cpu="0",mode="user"} 187.04
node_cpu_seconds_total{cpu="1",mode="idle"} 8403.82
node_cpu_seconds_total{cpu="1",mode="iowait"} 4.95
node_cpu_seconds_total{cpu="1",mode="irq"} 0
node_cpu_seconds_total{cpu="1",mode="nice"} 0
node_cpu_seconds_total{cpu="1",mode="softirq"} 16.75
node_cpu_seconds_total{cpu="1",mode="steal"} 0
node_cpu_seconds_total{cpu="1",mode="system"} 71.26
node_cpu_seconds_total{cpu="1",mode="user"} 190.27
100 74016 100 74016 0 0 5878k 0 --:--:-- --:--:-- --:--:-- 6023k
[root@master ~]# curl http://10.0.0.220:9100/metrics | grep node_load
% Total % Received % Xferd Average Speed Time Time Time Current
Dload Upload Total Spent Left Speed
0 0 0 0 0 0 0 0 --:--:-- --:--:-- --:--:-- 0# HELP node_loa
1 1m load average.
# TYPE node_load1 gauge
node_load1 0.2
# HELP node_load15 15m load average.
# TYPE node_load15 gauge
node_load15 0.22
# HELP node_load5 5m load average.
# TYPE node_load5 gauge
node_load5 0.2
100 74044 100 74044 0 0 8604k 0 --:--:-- --:--:-- --:--:-- 9038k
```
#### 2.Prometheus安装
创建sa账号对sa做rbac授权
```shell
# 创建一个sa账号monitor
[root@master ~]# kubectl create serviceaccount monitor -n monitor-sa
# 把sa账号monitor通过clusterrolebing绑定到clusterrole上
[root@master ~]# kubectl create clusterrolebinding monitor-clusterrolebinding -n monitor-sa --clusterrole=cluster-admin --serviceaccount=monitor-sa:monitor
```
创建prometheus数据存储目录
```shell
# 将prometheus调度到node-1节点
[root@node-1 ~]# mkdir /data && chmod 777 /data
```
创建一个configmap存储卷用来存放prometheus配置信息
```shell
[root@master ~]# vim prometheus-cfg.yaml
---
kind: ConfigMap
apiVersion: v1
metadata:
labels:
app: prometheus
name: prometheus-config
namespace: monitor-sa
data:
prometheus.yml: |
global:
scrape_interval: 15s
scrape_timeout: 10s
evaluation_interval: 1m
scrape_configs:
- job_name: 'kubernetes-node'
kubernetes_sd_configs:
- role: node
relabel_configs:
- source_labels: [__address__]
regex: '(.*):10250'
replacement: '${1}:9100'
target_label: __address__
action: replace
- action: labelmap
regex: __meta_kubernetes_node_label_(.+)
- job_name: 'kubernetes-node-cadvisor'
kubernetes_sd_configs:
- role: node
scheme: https
tls_config:
ca_file: /var/run/secrets/kubernetes.io/serviceaccount/ca.crt
bearer_token_file: /var/run/secrets/kubernetes.io/serviceaccount/token
relabel_configs:
- action: labelmap
regex: __meta_kubernetes_node_label_(.+)
- target_label: __address__
replacement: kubernetes.default.svc:443
- source_labels: [__meta_kubernetes_node_name]
regex: (.+)
target_label: __metrics_path__
replacement: /api/v1/nodes/${1}/proxy/metrics/cadvisor
- job_name: 'kubernetes-apiserver'
kubernetes_sd_configs:
- role: endpoints
scheme: https
tls_config:
ca_file: /var/run/secrets/kubernetes.io/serviceaccount/ca.crt
bearer_token_file: /var/run/secrets/kubernetes.io/serviceaccount/token
relabel_configs:
- source_labels: [__meta_kubernetes_namespace, __meta_kubernetes_service_name, __meta_kubernetes_endpoint_port_name]
action: keep
regex: default;kubernetes;https
- job_name: 'kubernetes-service-endpoints'
kubernetes_sd_configs:
- role: endpoints
relabel_configs:
- source_labels: [__meta_kubernetes_service_annotation_prometheus_io_scrape]
action: keep
regex: true
- source_labels: [__meta_kubernetes_service_annotation_prometheus_io_scheme]
action: replace
target_label: __scheme__
regex: (https?)
- source_labels: [__meta_kubernetes_service_annotation_prometheus_io_path]
action: replace
target_label: __metrics_path__
regex: (.+)
- source_labels: [__address__, __meta_kubernetes_service_annotation_prometheus_io_port]
action: replace
target_label: __address__
regex: ([^:]+)(?::\d+)?;(\d+)
replacement: $1:$2
- action: labelmap
regex: __meta_kubernetes_service_label_(.+)
- source_labels: [__meta_kubernetes_namespace]
action: replace
target_label: kubernetes_namespace
- source_labels: [__meta_kubernetes_service_name]
action: replace
target_label: kubernetes_name
```
创建:
```shell
[root@master ~]# kubectl apply -f prometheus-cfg.yaml
configmap/prometheus-config created
```
配置详解:
```shell
---
kind: ConfigMap
apiVersion: v1
metadata:
labels:
app: prometheus
name: prometheus-config
namespace: monitor-sa
data:
prometheus.yml: |
global:
scrape_interval: 15s #采集目标主机监控据的时间间隔
scrape_timeout: 10s # 数据采集超时时间默认10s
evaluation_interval: 1m #触发告警检测的时间默认是1m
scrape_configs: # 配置数据源称为target每个target用job_name命名。又分为静态配置和服务发现
- job_name: 'kubernetes-node'
kubernetes_sd_configs: # 使用的是k8s的服务发现
- role: node # 使用node角色它使用默认的kubelet提供的http端口来发现集群中每个node节点
relabel_configs: # 重新标记
- source_labels: [__address__] # 配置的原始标签,匹配地址
regex: '(.*):10250' #匹配带有10250端口的url
replacement: '${1}:9100' #把匹配到的ip:10250的ip保留
target_label: __address__ #新生成的url是${1}获取到的ip:9100
action: replace # 动作替换
- action: labelmap
regex: __meta_kubernetes_node_label_(.+) #匹配到下面正则表达式的标签会被保留,如果不做regex正则的话默认只是会显示instance标签
- job_name: 'kubernetes-node-cadvisor' # 抓取cAdvisor数据是获取kubelet上/metrics/cadvisor接口数据来获取容器的资源使用情况
kubernetes_sd_configs:
- role: node
scheme: https
tls_config:
ca_file: /var/run/secrets/kubernetes.io/serviceaccount/ca.crt
bearer_token_file: /var/run/secrets/kubernetes.io/serviceaccount/token
relabel_configs:
- action: labelmap # 把匹配到的标签保留
regex: __meta_kubernetes_node_label_(.+) #保留匹配到的具有__meta_kubernetes_node_label的标签
- target_label: __address__ # 获取到的地址__address__="192.168.40.180:10250"
replacement: kubernetes.default.svc:443 # 把获取到的地址替换成新的地址kubernetes.default.svc:443
- source_labels: [__meta_kubernetes_node_name]
regex: (.+) # 把原始标签中__meta_kubernetes_node_name值匹配到
target_label: __metrics_path__ #获取__metrics_path__对应的值
replacement: /api/v1/nodes/${1}/proxy/metrics/cadvisor
# 把metrics替换成新的值api/v1/nodes/k8s-master1/proxy/metrics/cadvisor
# ${1}是__meta_kubernetes_node_name获取到的值
# 新的url就是https://kubernetes.default.svc:443/api/v1/nodes/k8s-master1/proxy/metrics/cadvisor
- job_name: 'kubernetes-apiserver'
kubernetes_sd_configs:
- role: endpoints # 使用k8s中的endpoint服务发现采集apiserver 6443端口获取到的数据
scheme: https
tls_config:
ca_file: /var/run/secrets/kubernetes.io/serviceaccount/ca.crt
bearer_token_file: /var/run/secrets/kubernetes.io/serviceaccount/token
relabel_configs:
- source_labels: [__meta_kubernetes_namespace, __meta_kubernetes_service_name, __meta_kubernetes_endpoint_port_name]
# endpoint这个对象的名称空间,endpoint对象的服务名,exnpoint的端口名称
action: keep # 采集满足条件的实例,其他实例不采集
regex: default;kubernetes;https #正则匹配到的默认空间下的service名字是kubernetes协议是https的endpoint类型保留下来
- job_name: 'kubernetes-service-endpoints'
kubernetes_sd_configs:
- role: endpoints
relabel_configs:
- source_labels: [__meta_kubernetes_service_annotation_prometheus_io_scrape]
action: keep
regex: true
# 重新打标仅抓取到的具有 "prometheus.io/scrape: true" 的annotation的端点意思是说如果某个service具有prometheus.io/scrape = true annotation声明则抓取annotation本身也是键值结构所以这里的源标签设置为键而regex设置值true当值匹配到regex设定的内容时则执行keep动作也就是保留其余则丢弃。
- source_labels: [__meta_kubernetes_service_annotation_prometheus_io_scheme]
action: replace
target_label: __scheme__
regex: (https?)
# 重新设置scheme匹配源标签__meta_kubernetes_service_annotation_prometheus_io_scheme也就是prometheus.io/scheme annotation如果源标签的值匹配到regex则把值替换为__scheme__对应的值。
- source_labels: [__meta_kubernetes_service_annotation_prometheus_io_path]
action: replace
target_label: __metrics_path__
regex: (.+)
# 应用中自定义暴露的指标也许你暴露的API接口不是/metrics这个路径那么你可以在这个POD对应的service中做一个"prometheus.io/path = /mymetrics" 声明上面的意思就是把你声明的这个路径赋值给__metrics_path__其实就是让prometheus来获取自定义应用暴露的metrices的具体路径不过这里写的要和service中做好约定如果service中这样写 prometheus.io/app-metrics-path: '/metrics' 那么你这里就要__meta_kubernetes_service_annotation_prometheus_io_app_metrics_path这样写。
- source_labels: [__address__, __meta_kubernetes_service_annotation_prometheus_io_port]
action: replace
target_label: __address__
regex: ([^:]+)(?::\d+)?;(\d+)
replacement: $1:$2
# 暴露自定义的应用的端口就是把地址和你在service中定义的 "prometheus.io/port = <port>" 声明做一个拼接然后赋值给__address__这样prometheus就能获取自定义应用的端口然后通过这个端口再结合__metrics_path__来获取指标如果__metrics_path__值不是默认的/metrics那么就要使用上面的标签替换来获取真正暴露的具体路径。
- action: labelmap #保留下面匹配到的标签
regex: __meta_kubernetes_service_label_(.+)
- source_labels: [__meta_kubernetes_namespace]
action: replace # 替换__meta_kubernetes_namespace变成kubernetes_namespace
target_label: kubernetes_namespace
- source_labels: [__meta_kubernetes_service_name]
action: replace
target_label: kubernetes_name
```
通过deployment部署prometheus
```shell
[root@master ~]# cat prometheus-deploy.yaml
---
apiVersion: apps/v1
kind: Deployment
metadata:
name: prometheus-server
namespace: monitor-sa
labels:
app: prometheus
spec:
replicas: 1
selector:
matchLabels:
app: prometheus
component: server
#matchExpressions:
#- {key: app, operator: In, values: [prometheus]}
#- {key: component, operator: In, values: [server]}
template:
metadata:
labels:
app: prometheus
component: server
annotations:
prometheus.io/scrape: 'false'
spec:
nodeName: node-1 # 指定pod调度到哪个节点上
serviceAccountName: monitor
containers:
- name: prometheus
image: prom/prometheus:v2.2.1
#image: 10.0.0.230/xingdian/prometheus:v2.2.1
imagePullPolicy: IfNotPresent
command:
- prometheus
- --config.file=/etc/prometheus/prometheus.yml
- --storage.tsdb.path=/prometheus # 数据存储目录
- --storage.tsdb.retention=720h # 数据保存时长
- --web.enable-lifecycle # 开启热加载
ports:
- containerPort: 9090
protocol: TCP
volumeMounts:
- mountPath: /etc/prometheus/prometheus.yml
name: prometheus-config
subPath: prometheus.yml
- mountPath: /prometheus/
name: prometheus-storage-volume
volumes:
- name: prometheus-config
configMap:
name: prometheus-config
items:
- key: prometheus.yml
path: prometheus.yml
mode: 0644
- name: prometheus-storage-volume
hostPath:
path: /data
type: Directory
```
创建:
```shell
[root@master ~]# kubectl apply -f prometheus-deploy.yaml
deployment.apps/prometheus-server created
```
查看:
```shell
[root@master ~]# kubectl get pods -o wide -n monitor-sa
NAME READY STATUS RESTARTS AGE IP NODE NOMINATED NODE READINESS GATES
prometheus-server-59cb5d648-bxwrb 1/1 Running 0 14m 10.244.2.100 node-1 <none> <none>
```
#### 3.prometheus pod创建service
```shell
[root@master ~]# cat prometheus-svc.yaml
apiVersion: v1
kind: Service
metadata:
name: prometheus
namespace: monitor-sa
labels:
app: prometheus
spec:
type: NodePort
ports:
- port: 9090
targetPort: 9090
protocol: TCP
selector:
app: prometheus
component: server
```
创建:
```shell
[root@master ~]# kubectl apply -f prometheus-svc.yaml
service/prometheus created
```
查看service在物理机映射的端口
```shell
[root@master ~]# kubectl get svc -n monitor-sa
NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
prometheus NodePort 10.106.61.80 <none> 9090:32169/TCP 32m
```
#### 4.web界面查看
![image-20220602011956600](%E5%9F%BA%E4%BA%8Ekubernetes%E9%83%A8%E7%BD%B2Prometheus%E5%92%8CGrafana.assets/image-20220602011956600.png)
![image-20220602012012382](%E5%9F%BA%E4%BA%8Ekubernetes%E9%83%A8%E7%BD%B2Prometheus%E5%92%8CGrafana.assets/image-20220602012012382.png)
#### 5.Prometheus热加载
```shell
# 为了每次修改配置文件可以热加载prometheus也就是不停止prometheus就可以使配置生效想要使配置生效可用如下热加载命令
[root@master ~]# kubectl get pods -n monitor-sa -o wide -l app=prometheus
NAME READY STATUS RESTARTS AGE IP NODE NOMINATED NODE READINESS GATES
prometheus-server-689fb8cdbc-kcsw2 1/1 Running 0 5m39s 10.244.36.70 k8s-node1 <none> <none>
# 想要使配置生效可用如下命令热加载:
[root@master ~]# curl -X POST http://10.244.36.70:9090/-/reload
# 查看log
[root@master ~]# kubectl logs -n monitor-sa prometheus-server-689fb8cdbc-kcsw2
```
注意:
```shell
# 热加载速度比较慢可以暴力重启prometheus如修改上面的prometheus-cfg.yaml文件之后可执行如下强制删除
[root@master ~]# kubectl delete -f prometheus-cfg.yaml
[root@master ~]# kubectl delete -f prometheus-deploy.yaml
# 然后再通过apply更新
[root@master ~]# kubectl apply -f prometheus-cfg.yaml
[root@master ~]# kubectl apply -f prometheus-deploy.yaml
#注意:线上最好热加载,暴力删除可能造成监控数据的丢失
```
## 三Grafana的部署
#### 1.Grafana介绍
Grafana是一个跨平台的开源的度量分析和可视化工具可以将采集的数据可视化的展示并及时通知给告警接收方
它主要有以下六大特点:
1展示方式快速灵活的客户端图表面板插件有许多不同方式的可视化指标和日志官方库中具有丰富的仪表盘插件比如热图、折线图、图表等多种展示方式
2数据源GraphiteInfluxDBOpenTSDBPrometheusElasticsearchCloudWatch和KairosDB等
3通知提醒以可视方式定义最重要指标的警报规则Grafana将不断计算并发送通知在数据达到阈值时通过Slack、PagerDuty等获得通知
4混合展示在同一图表中混合使用不同的数据源可以基于每个查询指定数据源甚至自定义数据源
5注释使用来自不同数据源的丰富事件注释图表将鼠标悬停在事件上会显示完整的事件元数据和标记
#### 2.Grafana安装
```shell
[root@master prome]# cat grafana.yaml
apiVersion: apps/v1
kind: Deployment
metadata:
name: monitoring-grafana
namespace: kube-system
spec:
replicas: 1
selector:
matchLabels:
task: monitoring
k8s-app: grafana
template:
metadata:
labels:
task: monitoring
k8s-app: grafana
spec:
containers:
- name: grafana
image: 10.0.0.230/xingdian/heapster-grafana-amd64:v5.0.4
#heleicool/heapster-grafana-amd64:v5.0.4
ports:
- containerPort: 3000
protocol: TCP
volumeMounts:
- mountPath: /etc/ssl/certs
name: ca-certificates
readOnly: true
- mountPath: /var
name: grafana-storage
env:
- name: INFLUXDB_HOST
value: monitoring-influxdb
- name: GF_SERVER_HTTP_PORT
value: "3000"
# The following env variables are required to make Grafana accessible via
# the kubernetes api-server proxy. On production clusters, we recommend
# removing these env variables, setup auth for grafana, and expose the grafana
# service using a LoadBalancer or a public IP.
- name: GF_AUTH_BASIC_ENABLED
value: "false"
- name: GF_AUTH_ANONYMOUS_ENABLED
value: "true"
- name: GF_AUTH_ANONYMOUS_ORG_ROLE
value: Admin
- name: GF_SERVER_ROOT_URL
# If you're only using the API Server proxy, set this value instead:
# value: /api/v1/namespaces/kube-system/services/monitoring-grafana/proxy
value: /
volumes:
- name: ca-certificates
hostPath:
path: /etc/ssl/certs
- name: grafana-storage
emptyDir: {}
---
apiVersion: v1
kind: Service
metadata:
labels:
# For use as a Cluster add-on (https://github.com/kubernetes/kubernetes/tree/master/cluster/addons)
# If you are NOT using this as an addon, you should comment out this line.
kubernetes.io/cluster-service: 'true'
kubernetes.io/name: monitoring-grafana
name: monitoring-grafana
namespace: kube-system
spec:
# In a production setup, we recommend accessing Grafana through an external Loadbalancer
# or through a public IP.
# type: LoadBalancer
# You could also use NodePort to expose the service at a randomly-generated port
# type: NodePort
ports:
- port: 80
targetPort: 3000
selector:
k8s-app: grafana
type: NodePort
```
创建:
```shell
[root@master prome]# kubectl apply -f grafana.yaml
deployment.apps/monitoring-grafana created
service/monitoring-grafana created
```
查看:
```shell
[root@master prome]# kubectl get pods -n kube-system -l task=monitoring -o wide
NAME READY STATUS RESTARTS AGE IP NODE NOMINATED NODE READINESS GATES
monitoring-grafana-7c5c6c7486-rbt62 1/1 Running 0 9s 10.244.1.83 node-3 <none> <none>
```
```shell
[root@master prome]# kubectl get svc -n kube-system | grep grafana
monitoring-grafana NodePort 10.101.77.194 <none> 80:30919/TCP 76s
```
## 四配置Grafana
浏览器访问:
![image-20220602013222284](%E5%9F%BA%E4%BA%8Ekubernetes%E9%83%A8%E7%BD%B2Prometheus%E5%92%8CGrafana.assets/image-20220602013222284.png)
添加数据源:
![image-20220602013322234](%E5%9F%BA%E4%BA%8Ekubernetes%E9%83%A8%E7%BD%B2Prometheus%E5%92%8CGrafana.assets/image-20220602013322234.png)
指定Prometheus地址
![image-20220602013441712](%E5%9F%BA%E4%BA%8Ekubernetes%E9%83%A8%E7%BD%B2Prometheus%E5%92%8CGrafana.assets/image-20220602013441712.png)
导入监控模板:
![image-20220602013943317](%E5%9F%BA%E4%BA%8Ekubernetes%E9%83%A8%E7%BD%B2Prometheus%E5%92%8CGrafana.assets/image-20220602013943317.png)
![image-20220602014027197](%E5%9F%BA%E4%BA%8Ekubernetes%E9%83%A8%E7%BD%B2Prometheus%E5%92%8CGrafana.assets/image-20220602014027197.png)
注意:
官方下载监控模板https://grafana.com/dashboards?dataSource=prometheus&search=kubernetes
![image-20220602014152927](%E5%9F%BA%E4%BA%8Ekubernetes%E9%83%A8%E7%BD%B2Prometheus%E5%92%8CGrafana.assets/image-20220602014152927.png)
![image-20220602014212551](%E5%9F%BA%E4%BA%8Ekubernetes%E9%83%A8%E7%BD%B2Prometheus%E5%92%8CGrafana.assets/image-20220602014212551.png)
展示:
![image-20220602014306247](%E5%9F%BA%E4%BA%8Ekubernetes%E9%83%A8%E7%BD%B2Prometheus%E5%92%8CGrafana.assets/image-20220602014306247.png)
![image-20220602014321106](%E5%9F%BA%E4%BA%8Ekubernetes%E9%83%A8%E7%BD%B2Prometheus%E5%92%8CGrafana.assets/image-20220602014321106.png)
![image-20220602014337431](%E5%9F%BA%E4%BA%8Ekubernetes%E9%83%A8%E7%BD%B2Prometheus%E5%92%8CGrafana.assets/image-20220602014337431.png)