kubernetes學習筆記之十二：資源指標API及自定義指標API

阿新 • • 發佈：2018-12-22

第一章、前言

以前是用heapster來收集資源指標才能看，現在heapster要廢棄了
從1.8以後引入了資源api指標監視
資源指標：metrics-server(核心指標)
自定義指標：prometheus，k8s-prometheus-adapter(將Prometheus採集的資料轉換為指標格式)
    k8s的中的prometheus需要k8s-prometheus-adapter轉換一下才可以使用
新一代架構：
    核心指標流水線：
        kubelet，metrics 
-service以及API service提供api組成；cpu累計使用率，記憶體實時使用率，pod的資源佔用率和容器磁碟佔用率；
    監控流水線：
        用於從系統收集各種指標資料並提供終端使用者，儲存系統以及HPA，他們包括核心指標以及很多非核心指標，非核心指標本身不能被k8s解析

第二章、安裝部署metrics-server

1、下載yaml檔案,並安裝

專案地址：https://github.com/kubernetes/kubernetes/tree/master/cluster/addons/metrics-server

，選擇與版本對應的分支，我的是v1.10.0,所以這裡我選擇v1.10.0分支

[ro[email protected]_01 manifests]# mkdir metrics-server
[[email protected]-master_01 manifests]# cd metrics-server
[[email protected]-master_01 metrics-server]# for file in auth-delegator.yaml auth-reader.yaml metrics-apiservice.yaml metrics-server-deployment.yaml metrics-server-service.yaml resource-reader.yaml;do 
 wget https://raw.githubusercontent.com/kubernetes/kubernetes/v1.10.0/cluster/addons/metrics-server/$file;done   #記住，下載raw格式的檔案
[[email protected]_01 metrics-server]# grep image: ./*  #檢視使用的映象，如果可以科學上網，那麼忽略，如果不可用那麼需要提前下載，通過修改配置檔案或修改映象的名稱的方式載入映象，映象可以到阿里雲上去搜索
./metrics-server-deployment.yaml:        image: k8s.gcr.io/metrics-server-amd64:v0.2.1
./metrics-server-deployment.yaml:        image: k8s.gcr.io/addon-resizer:1.8.1
[[email protected]_01 ~]# docker pull registry.cn-hangzhou.aliyuncs.com/criss/addon-resizer:1.8.1  #手動在所有的node節點上下載映象，注意版本號沒有v
[[email protected]_01 ~]# docker pull registry.cn-hangzhou.aliyuncs.com/k8s-kernelsky/metrics-server-amd64:v0.2.1
[[email protected]_01 metrics-server]# grep image: metrics-server-deployment.yaml
        image: registry.cn-hangzhou.aliyuncs.com/k8s-kernelsky/metrics-server-amd64:v0.2.1
        image: registry.cn-hangzhou.aliyuncs.com/criss/addon-resizer:1.8.1
[[email protected]_01 metrics-server]# kubectl apply -f .
[[email protected]_01 metrics-server]# kubectl get pod -n kube-system

2、驗證

[[email protected] ~]# kubectl api-versions |grep metrics
metrics.k8s.io/v1beta1
[[email protected] ~]# kubectl proxy --port=8080  #重新開啟一個終端，啟動代理功能
[[email protected]-master_01 metrics-server]# curl http://localhost:8080/apis/metrics.k8s.io/v1beta1 #檢視這個資源組包含哪些元件
[[email protected]_01 metrics-server]# curl http://localhost:8080/apis/metrics.k8s.io/v1beta1/pods  #可能需要等待一會在會有資料
[[email protected]_01 metrics-server]# curl http://localhost:8080/apis/metrics.k8s.io/v1beta1/nodes
[[email protected] ~]# kubectl top node
NAME           CPU(cores)   CPU%      MEMORY(bytes)   MEMORY%   
k8s-master01   176m         4%        3064Mi          39%       
k8s-node01     62m          1%        4178Mi          54%       
k8s-node02     65m          1%        2141Mi          27%       
[[email protected]-node01 ~]# kubectl top pods
NAME                CPU(cores)   MEMORY(bytes)   
node-affinity-pod   0m           1Mi

3.注意事項

1.#在更新的版本中，如v1.11及以上會出現問題，這是因為metric-service預設從kubernetes的summary_api中獲取資料，而summary_api預設使用10255埠來獲
取資料，但是10255是一個http協議的埠，可能官方認為http協議不安全所以封禁了10255埠改為使用10250埠，而10250是一個https協議埠，所以我們需要修改一下連線方式：
由  - --source=kubernetes.summary_api:''
修改為  - --source=kubernetes.summary_api:https://kubernetes.default?kubeletHttps=true&kubeletPort=10250&insecure-true  #表示雖然我使用https協議來通訊，並且埠也是10250，但是如果證書不能認證依然可以通過非安全不加密的方式來通訊
[[email protected] deploy]# grep source=kubernetes  metrics-server-deployment.yaml 
2.[[email protected] deploy]# grep nodes/stats  resource-reader.yaml #在新的版本中，授權文內沒有 node/stats 的許可權，需要手動去新增
[[email protected] deploy]# cat resource-reader.yaml 
apiVersion: rbac.authorization.k8s.io/v1
kind: ClusterRole
metadata:
  name: system:metrics-server
rules:
- apiGroups:
  - ""
  resources:
  - pods
  - nodes
  - nodes/stats  #新增這一行
  - namespaces
3.在1.12.3版本中測試發現，需要進行如下修改才能成功部署(許可權依然需要修改，其他版本暫未測試)
[[email protected] metrics-server]# vim metrics-server-deployment.yaml
command:   #metrics-server命令引數修改為如下引數
  - /metrics-server
  - --metric-resolution=30s
  - --kubelet-port=10250
  - --kubelet-insecure-tls
  - --kubelet-preferred-address-types=InternalIP
command:    #metrics-server-nanny 命令引數修改為如下引數
  - /pod_nanny
  - --config-dir=/etc/config
  - --cpu=40m
  - --extra-cpu=0.5m
  - --memory=40Mi
  - --extra-memory=4Mi
  - --threshold=5
  - --deployment=metrics-server-v0.3.1
  - --container=metrics-server
  - --poll-period=300000
  - --estimator=exponential

第三章、安裝部署prometheus

專案地址：https://github.com/kubernetes/kubernetes/tree/master/cluster/addons/prometheus(由於prometheus只有v1.11.0及以上才有，所有我選擇v1.11.0來部署)

1.下載yaml檔案及部署前操作

[[email protected] ~]# cd /mnt/
[[email protected]-node01 mnt]# git clone https://github.com/kubernetes/kubernetes.git  #我嫌麻煩就直接克隆kubernetes整個專案了
[[email protected] mnt]# cd kubernetes/cluster/addons/prometheus/
[[email protected]-node01 prometheus]# git checkout v1.11.0
[[email protected]-node01 prometheus]# cd ..
[[email protected]-node01 addons]# cp -r prometheus /root/manifests/
[[email protected]-node01 manifests]# cd prometheus/
[[email protected]-node01 prometheus]# grep -w  "namespace: kube-system" ./*   #預設prometheus使用的是kube-system名稱空間，我們把它單獨部署到一個名稱空間中，方便之後的管理
./alertmanager-configmap.yaml:  namespace: kube-system
......
[r[email protected] prometheus]# sed  -i 's/namespace: kube-system/namespace\: k8s-monitor/g' ./* 
[[email protected] prometheus]# grep storage: ./*   #安裝需要兩個pv，等下我們需要建立一下
./alertmanager-pvc.yaml:      storage: "2Gi"
./prometheus-statefulset.yaml:          storage: "16Gi"
[[email protected] prometheus]# cat pv.yaml #注意第二pv的storageClassName
apiVersion: v1
kind: PersistentVolume
metadata:
  name: alertmanager  
spec:
  capacity: 
    storage: 5Gi
  accessModes: 
    - ReadWriteOnce 
    - ReadWriteMany
  persistentVolumeReclaimPolicy: Recycle
  nfs:
    path: /data/volumes/v1
    server: 172.16.150.158
---
apiVersion: v1
kind: PersistentVolume
metadata: 
  name: standard
spec:
  capacity: 
    storage: 25Gi
  accessModes:
    - ReadWriteOnce
  persistentVolumeReclaimPolicy: Recycle
  storageClassName: standard   #storageClassName與prometheus-statefulset.yaml中volumeClaimTemplates下定義的需要保持一致
  nfs:
    path: /data/volumes/v2
    server: 172.16.150.158
[[email protected] prometheus]# kubectl create namespace k8s-monitor
[[email protected] prometheus]# mkdir node-exporter kube-state-metrics alertmanager prometheus #將每個元件單獨放入一個目錄中，方便部署及管理
[[email protected] prometheus]# mv node-exporter-* node-exporter
[[email protected] prometheus]# mv alertmanager-* alertmanager
[[email protected] prometheus]# mv kube-state-metrics-* kube-state-metrics
[[email protected] prometheus]# mv prometheus-* prometheus

2.安裝node-exporter(用於收集節點的資料指標)

[[email protected] prometheus]# grep -r image:  node-exporter/*
node-exporter/node-exporter-ds.yml:          image: "prom/node-exporter:v0.15.2"   #非官方映象，不能科學上網的也可以下載，所以不需要提前下載
[[email protected] prometheus]# kubectl apply -f node-exporter/
daemonset.extensions "node-exporter" created
service "node-exporter" created
[[email protected] prometheus]# kubectl get pod -n k8s-monitor 
NAME                  READY     STATUS    RESTARTS   AGE
node-exporter-l5zdw   1/1       Running   0          1m
node-exporter-vwknx   1/1       Running   0          1m

3.安裝prometheus

[[email protected]_01 prometheus]# kubectl apply -f pv.yaml 
persistentvolume "alertmanager" configured
persistentvolume "standard" created
[[email protected]-master_01 prometheus]# kubectl get pv
NAME              CAPACITY   ACCESS MODES   RECLAIM POLICY   STATUS      CLAIM     STORAGECLASS   REASON    AGE
alertmanager      5Gi        RWO,RWX        Recycle          Available                                      9s

standard          25Gi       RWO            Recycle          Available                                      9s
[[email protected]-node01 prometheus]# grep -i image prometheus/*  #檢視映象是否需要下載
[[email protected] prometheus]# vim prometheus-service.yaml   #預設prometheus的service埠型別為ClusterIP，為了可以叢集外訪問，修改為NodePort
...
  type: NodePort
  ports:
    - name: http
      port: 9090
      protocol: TCP
      targetPort: 9090
      nodePort: 30090
...
[[email protected] prometheus]# kubectl apply -f prometheus/
[[email protected] prometheus]# kubectl get pod -n k8s-monitor 
NAME                  READY     STATUS    RESTARTS   AGE
node-exporter-l5zdw   1/1       Running   0          24m
node-exporter-vwknx   1/1       Running   0          24m
prometheus-0          2/2       Running   0          1m
[[email protected] prometheus]# kubectl get svc -n k8s-monitor 
NAME            TYPE        CLUSTER-IP    EXTERNAL-IP   PORT(S)          AGE
node-exporter   ClusterIP   None          <none>        9100/TCP         25m
prometheus      NodePort    10.96.9.121   <none>        9090:30090/TCP   22m
[[email protected]_01 prometheus]# kubectl get pv
NAME           CAPACITY   ACCESS MODES   RECLAIM POLICY   STATUS      CLAIM                                      STORAGECLASS   REASON    AGE
alertmanager   5Gi        RWO,RWX        Recycle          Available                                                                       1h
standard       25Gi       RWO            Recycle          Bound       k8s-monitor/prometheus-data-prometheus-0   standard                 1h

訪問prometheus(node節點IP：埠)

4.部署metrics介面卡(將prometheus資料轉換為k8s可以識別的資料)

[[email protected] kube-state-metrics]# grep image: ./*
./kube-state-metrics-deployment.yaml:        image: quay.io/coreos/kube-state-metrics:v1.3.0
./kube-state-metrics-deployment.yaml:        image: k8s.gcr.io/addon-resizer:1.7
[[email protected] ~]#  docker pull registry.cn-hangzhou.aliyuncs.com/ccgg/addon-resizer:1.7
[[email protected] kube-state-metrics]# vim kube-state-metrics-deployment.yaml   #修改映象地址
[[email protected] kube-state-metrics]# kubectl apply -f kube-state-metrics-deployment.yaml
deployment.extensions "kube-state-metrics" configured
[[email protected] kube-state-metrics]# kubectl get pod -n k8s-monitor 
NAME                                  READY     STATUS    RESTARTS   AGE
kube-state-metrics-54849b96b4-dmqtk   2/2       Running   0          23s
node-exporter-l5zdw                   1/1       Running   0          2h
node-exporter-vwknx                   1/1       Running   0          2h
prometheus-0                          2/2       Running   0          1h

5.部署k8s-prometheus-adapter(將資料輸出為一個API服務)

專案地址：https://github.com/DirectXMan12/k8s-prometheus-adapter

[[email protected] ~]# cd /etc/kubernetes/pki/
[[email protected] pki]#(umask 077; openssl genrsa -out serving.key 2048)
[[email protected] pki]#openssl req -new -key serving.key -out serving.csr -subj "/CN=serving"  #CN必須為serving
[[email protected] pki]#openssl x509 -req -in serving.csr -CA ./ca.crt -CAkey ./ca.key  -CAcreateserial -out serving.crt -days 3650
[[email protected] pki]# kubectl create secret generic cm-adapter-serving-certs --from-file=serving.crt=./serving.crt --from-file=serving.key=./serving.key -n k8s-monitor #證書名稱必須為cm-adapter-serving-certs
[[email protected] pki]#kubectl get secret  -n k8s-monitor
[[email protected] pki]# cd
[[email protected] ~]# git clone https://github.com/DirectXMan12/k8s-prometheus-adapter.git
[[email protected] ~]# cd k8s-prometheus-adapter/deploy/manifests/
[[email protected]-node01 manifests]# grep namespace: ./* #處理role-binding之外的namespace的名稱改為k8s-monitor
[[email protected] manifests]# grep image: ./* #映象不需要下載
[[email protected] ~]# sed -i 's/namespace\: custom-metrics/namespace\: k8s-monitor/g' ./*   #rolebinding的不要替換
[[email protected] ~]# kubectl apply -f ./
[[email protected] ~]# kubectl get pod -n k8s-monitor
[[email protected] ~]#kubectl get svc -n k8s-monitor
kubectl api-versions |grep custom

第四章、部署prometheus+grafana

[[email protected] ~]# wget https://raw.githubusercontent.com/kubernetes-retired/heapster/master/deploy/kube-config/influxdb/grafana.yaml #找不到grafana的yaml檔案，所以到heapster裡面掏了一個下來用用
[[email protected] ~]#egrep -i "influxdb|namespace|nodeport" grafana.yaml  #註釋掉influxdb環境變數，修改namespace及port型別
[[email protected]-master01 ~]#kubectl apply -f grafana.yaml
[[email protected]-master01 ~]#kubectl get svc  -n k8s-monitor
[[email protected]-master01 ~]#kubectl get pod -n k8s-monitor

登入grafana，並修改資料來源

配置資料來源

點選右側的Dashborads，可以匯入grafana自帶的prometheus的模板

回到home下，下拉選擇對應的模板檢視資料

例如：

但是，grafana自帶的模板和資料有些不匹配，我們可以去grafana官網去下載應用於k8s使用的模板，地址為：https://grafana.com/dashboards

訪問grafana官網搜尋k8s相關模板，有時搜尋框點選沒有反應，可以直接在URL後面加上搜索內容即可

我們選擇kubernetes cluster(prometheus)作為測試

點選需要下載的模板，並下載json檔案

下載完成後，匯入檔案

選擇上傳檔案

匯入後選擇資料來源

匯入後展示的介面

第五章、實現HPA　　

1、使用v1版本測試

[[email protected] alertmanager]# kubectl api-versions |grep autoscaling
autoscaling/v1
autoscaling/v2beta1
[[email protected]-master01 manifests]# cat deploy-demon.yaml
apiVersion: v1
kind: Service
metadata:
  name: myapp
  namespace: default
spec:
  selector:
    app: myapp
  type: NodePort
  ports:
  - name: http
    port: 80
    targetPort: 80
    nodePort: 32222

---
apiVersion: apps/v1
kind: Deployment
metadata: 
  name: myapp-deploy
spec:
  replicas: 2
  selector: 
    matchLabels:
      app: myapp
  template:
    metadata:
      labels:
        app: myapp
    spec:
      containers:
      - name: myapp
        image: ikubernetes/myapp:v2
        ports:
        - name: httpd
          containerPort: 80
        resources:
          requests:
            memory: "64Mi"
            cpu: "100m"
          limits:
            memory: "128Mi"
            cpu: "200m"
[[email protected]-master01 manifests]# kubectl get svc
NAME         TYPE        CLUSTER-IP      EXTERNAL-IP   PORT(S)             AGE
kubernetes   ClusterIP   10.96.0.1       <none>        443/TCP             47d
my-nginx     NodePort    10.104.13.148   <none>        80:32008/TCP        19d
myapp        NodePort    10.100.76.180   <none>        80:32222/TCP        16s
tomcat       ClusterIP   10.106.222.72   <none>        8080/TCP,8009/TCP   19d
[[email protected]-master01 manifests]# kubectl get pod
NAME                            READY     STATUS    RESTARTS   AGE
myapp-deploy-5db497dbfb-h7zcb   1/1       Running   0          16s
myapp-deploy-5db497dbfb-tvsf5   1/1       Running   0          16s

測試

[[email protected] manifests]# kubectl autoscale deployment myapp-deploy --min=1 --max=8 --cpu-percent=60
deployment.apps "myapp-deploy" autoscaled
[[email protected]-master01 manifests]# kubectl get hpa
NAME           REFERENCE                 TARGETS         MINPODS   MAXPODS   REPLICAS   AGE
myapp-deploy   Deployment/myapp-deploy   <unknown>/60%   1         8         0          22s
[[email protected]-master01 pod-dir]# yum install http-tools -y
[[email protected]-master01 pod-dir]# ab -c 1000 -n 5000000 http://172.16.150.213:32222/index.html
[[email protected] ~]# kubectl describe hpa 
Name:                                                  myapp-deploy
Namespace:                                             default
Labels:                                                <none>
Annotations:                                           <none>
CreationTimestamp:                                     Sun, 16 Dec 2018 20:34:41 +0800
Reference:                                             Deployment/myapp-deploy
Metrics:                                               ( current / target )
  resource cpu on pods  (as a percentage of request):  178% (178m) / 60%
Min replicas:                                          1
Max replicas:                                          8
Conditions:
  Type            Status  Reason            Message
  ----            ------  ------            -------
  AbleToScale     False   BackoffBoth       the time since the previous scale is still within both the downscale and upscale forbidden windows
  ScalingActive   True    ValidMetricFound  the HPA was able to successfully calculate a replica count from cpu resource utilization (percentage of request)
  ScalingLimited  True    ScaleUpLimit      the desired replica count is increasing faster than the maximum scale rate
Events:
  Type    Reason             Age   From                       Message
  ----    ------             ----  ----                       -------
  Normal  SuccessfulRescale  19m   horizontal-pod-autoscaler  New size: 1; reason: All metrics below target
  Normal  SuccessfulRescale  2m    horizontal-pod-autoscaler  New size: 2; reason: cpu resource utilization (percentage of request) above target
[[email protected]-master01 ~]# kubectl get pod
NAME                            READY     STATUS    RESTARTS   AGE
myapp-deploy-5db497dbfb-6kssf   1/1       Running   0          2m
myapp-deploy-5db497dbfb-h7zcb   1/1       Running   0          24m
[[email protected]-master01 ~]# kubectl get hpa
NAME           REFERENCE                 TARGETS    MINPODS   MAXPODS   REPLICAS   AGE
myapp-deploy   Deployment/myapp-deploy   178%/60%   1         8         2          20m

2、使用v2beat1

[[email protected] pod-dir]# cat hpa-demo.yaml 
apiVersion: autoscaling/v2beta1
kind: HorizontalPodAutoscaler
metadata:
  name: myapp-hpa-v2
spec:
  scaleTargetRef:
    apiVersion: apps/v1
    kind: Deployment
    name: myapp-deploy
  minReplicas: 1
  maxReplicas: 10
  metrics:
  - type: Resource
    resource:
      name: cpu
      targetAverageUtilization: 55
  - type: Resource
    resource:
      name: memory
      targetAverageValue: 100Mi
[[email protected]-master01 pod-dir]# kubectl delete hpa myapp-deploy 
horizontalpodautoscaler.autoscaling "myapp-deploy" deleted
[[email protected]-master01 pod-dir]# kubectl apply -f hpa-demo.yaml 
horizontalpodautoscaler.autoscaling "myapp-hpa-v2" created
[[email protected]-master01 pod-dir]# kubectl get hpa
NAME           REFERENCE                 TARGETS                          MINPODS   MAXPODS   REPLICAS   AGE
myapp-hpa-v2   Deployment/myapp-deploy   <unknown>/100Mi, <unknown>/55%   1         10        0          6s

測試

[[email protected] ~]# kubectl describe hpa 
Name:                                                  myapp-hpa-v2
Namespace:                                             default
Labels:                                                <none>
Annotations:                                           kubectl.kubernetes.io/last-applied-configuration={"apiVersion":"autoscaling/v2beta1","kind":"HorizontalPodAutoscaler","metadata":{"annotations":{},"name":"myapp-hpa-v2","namespace":"default"},"spec":{...
CreationTimestamp:                                     Sun, 16 Dec 2018 21:07:25 +0800
Reference:                                             Deployment/myapp-deploy
Metrics:                                               ( current / target )
  resource memory on pods:                             1765376 / 100Mi
  resource cpu on pods  (as a percentage of request):  200% (200m) / 55%
Min replicas:                                          1
Max replicas:                                          10
Conditions:
  Type            Status  Reason              Message
  ----            ------  ------              -------
  AbleToScale     True    SucceededRescale    the HPA controller was able to update the target scale to 4
  ScalingActive   True    ValidMetricFound    the HPA was able to successfully calculate a replica count from cpu resource utilization (percentage of request)
  ScalingLimited  False   DesiredWithinRange  the desired count is within the acceptable range
Events:
  Type    Reason             Age   From                       Message
  ----    ------             ----  ----                       -------
  Normal  SuccessfulRescale  18s   horizontal-pod-autoscaler  New size: 4; reason: cpu resource utilization (percentage of request) above target
[[email protected]-master01 ~]# kubectl get pod
NAME                            READY     STATUS    RESTARTS   AGE
myapp-deploy-5db497dbfb-5n885   1/1       Running   0          26s
myapp-deploy-5db497dbfb-h7zcb   1/1       Running   0          40m
myapp-deploy-5db497dbfb-z2tqd   1/1       Running   0          26s
myapp-deploy-5db497dbfb-zkjhw   1/1       Running   0          26s
[[email protected]-master01 ~]# kubectl describe hpa 
Name:                                                  myapp-hpa-v2
Namespace:                                             default
Labels:                                                <none>
Annotations:                                           kubectl.kubernetes.io/last-applied-configuration={"apiVersion":"autoscaling/v2beta1","kind":"HorizontalPodAutoscaler","metadata":{"annotations":{},"name":"myapp-hpa-v2","namespace":"default"},"spec":{...
CreationTimestamp:                                     Sun, 16 Dec 2018 21:07:25 +0800
Reference:                                             Deployment/myapp-deploy
Metrics:                                               ( current / target )
  resource memory on pods:                             1765376 / 100Mi
  resource cpu on pods  (as a percentage of request):  0% (0) / 55%
Min replicas:                                          1
Max replicas:                                          10
Conditions:
  Type            Status  Reason              Message
  ----            ------  ------              -------
  AbleToScale     False   BackoffBoth         the time since the previous scale is still within both the downscale and upscale forbidden windows
  ScalingActive   True    ValidMetricFound    the HPA was able to successfully calculate a replica count from memory resource
  ScalingLimited  False   DesiredWithinRange  the desired count is within the acceptable range
Events:
  Type    Reason             Age   From                       Message
  ----    ------             ----  ----                       -------
  Normal  SuccessfulRescale  6m    horizontal-pod-autoscaler  New size: 4; reason: cpu resource utilization (percentage of request) above target
  Normal  SuccessfulRescale  34s   horizontal-pod-autoscaler  New size: 1; reason: All metrics below target
[[email protected]-master01 ~]# kubectl get pod
NAME                            READY     STATUS    RESTARTS   AGE
myapp-deploy-5db497dbfb-h7zcb   1/1       Running   0          46m

3.使用v2beat1測試自定義選項

[[email protected] pod-dir]# cat  ../deploy-demon-metrics.yaml
apiVersion: v1
kind: Service
metadata:
  name: myapp
  namespace: default
spec:
  selector:
    app: myapp
  type: NodePort
  ports:
  - name: http
    port: 80
    targetPort: 80
    nodePort: 32222

---
apiVersion: apps/v1
kind: Deployment
metadata: 
  name: myapp-deploy
spec:
  replicas: 2
  selector: 
    matchLabels:
      app: myapp
  template:
    metadata:
      labels:
        app: myapp
    spec:
      containers:
      - name: myapp
        image: ikubernetes/metrics-app  #測試映象
        ports:
        - name: httpd
          containerPort: 80
[[email protected]-master01 pod-dir]# kubectl apply -f deploy-demon-metrics.yaml
[[email protected]-master01 pod-dir]# cat hpa-custom.yaml 
apiVersion: autoscaling/v2beta1
kind: HorizontalPodAutoscaler
metadata:
  name: myapp-hpa-v2
spec:
  scaleTargetRef:
    apiVersion: apps/v1
    kind: Deployment
    name: myapp-deploy
  minReplicas: 1
  maxReplicas: 10
  metrics:
  - type: Pods   #注意型別
    pods:
      metricName: http_requests #容器中自定義的引數
      targetAverageValue: 800m  #m表示個數，即800個併發數
[[email protected]-master01 pod-dir]# kubectl apply -f hpa-custom.yaml 
[[email protected]-master01 pod-dir]# kubectl describe hpa myapp-hpa-v2 
Name:                       myapp-hpa-v2
Namespace:                  default
Labels:                     <none>
Annotations:                kubectl.kubernetes.io/last-applied-configuration={"apiVersion":"autoscaling/v2beta1","ks":{},"name":"myapp-hpa-v2","namespace":"default"},"spec":{...
CreationTimestamp:          Sun, 16 Dec 2018 22:09:32 +0800
Reference:                  Deployment/myapp-deploy
Metrics:                    ( current / target )
  "http_requests" on pods:  <unknown> / 800m
Min replicas:               1
Max replicas:               10
Events:                     <none>
[[email protected]-master01 pod-dir]# kubectl get hpa
NAME           REFERENCE                 TARGETS          MINPODS   MAXPODS   REPLICAS   AGE
myapp-hpa-v2   Deployment/myapp-deploy   <unknown>/800m   1         10        2          5m

測試：

#好像映象有點問題，待解決

kubernetes學習筆記之十二：資源指標API及自定義指標API

第一章、前言以前是用heapster來收集資源指標才能看，現在heapster要廢棄了從1.8以後引入了資源api指標監視資源指標：metrics-server(核心指標) 自定義指標：prometheus，k8s-prometheus-adapter(將Prometheus採集的資料轉換為指

kubernetes學習筆記之十一：kubernetes dashboard認證及分級授權

第一章、部署dashboard 作為Kubernetes的Web使用者介面，使用者可以通過Dashboard在Kubernetes叢集中部署容器化的應用，對應用進行問題處理和管理，並對叢集本身進行管理。通過Dashboard，使用者可以檢視叢集中應用的執行情況，同時也能夠基於Dashboard建立或修

Java Web 學習筆記之十二：JBoss RestEasy處理跨域OPTIONS請求方式

跨域請求問題前提前後端分離的B/S架構系統前後端獨立開發，後端採用JBoss restEasy 框架搭建restful服務後端程式碼開發完成並且部署在某一臺測試機上前端開發過程中，需

kubernetes學習筆記之十四：helm入門

1.Helm的簡介 Helm是Kubernetes的一個包管理工具，用來簡化Kubernetes應用的部署和管理。可以把Helm比作CentOS的yum工具。 Helm有如下幾個基本概念： Chart: 是Helm管理的安裝包，裡面包含需要部署的安裝包資源。可以把Chart比作CentOS yum使

【Unity 3D】學習筆記四十二：粒子特效

空間獲得 material package 一個 log 創建 spa mpi 粒子特效粒子特效的原理是將若幹粒子無規則的組合在一起。來模擬火焰，爆炸。水滴，霧氣等效果。要使用粒子特效首先要創建，在hierarchy視圖中點擊create——particle s

《SAS編程與數據挖掘商業案例》學習筆記之十二

style rename today 降序排序 cat list append span nod 本次重點在：sas數據集管理主要包含：包含數據集縱向拼接、轉置、排序、比較、復制、重命名、刪除等 1.append語句註：base數據集

PID控制器開發筆記之十二：模糊PID控制器的實現

poi 技術 last 全部 arr linear 數值 .com 數據　　在現實控制中，被控系統並非是線性時不變的，往往需要動態調整PID的參數，而模糊控制正好能夠滿足這一需求，所以在接下來的這一節我們將討論模糊PID控制器的相關問題。模糊PID控制器是將模糊算法與PI

kubernetes學習筆記之十:RBAC

第一章、RBAC介紹在Kubernetes中，授權有ABAC（基於屬性的訪問控制）、RBAC（基於角色的訪問控制）、Webhook、Node、AlwaysDeny（一直拒絕）和AlwaysAllow（一直允許）這6種模式。從1.6版本起，Kubernetes 預設啟用RBAC訪問控制策略。從1.8開

opencv學習筆記三十二：Haar特徵與積分影象

一、 Haar特徵定義 Haar特徵是基於“塊”的特徵，也被稱為矩形特徵。Haar特徵（模板）分為三類：邊緣特徵、線性特徵、中心特徵和對角線特徵。特徵模板內有白色和黑色兩種矩形，並定義該模板的特徵值為白色矩形畫素和減去黑色矩形畫素和。Haar特徵值反映了影象

opencv學習筆記四十二：稠密光流跟蹤

利用Gunnar Farneback演算法計算全域性性的稠密光流演算法（即影象上所有畫素點的光流都計算出來），由於要計算影象上所有點的光流，故計算耗時，速度慢稠密光流需要使用某種插值方法在比較容易跟蹤的畫素之間進行插值以解決那些運動不明確的畫素 calcOpticalF

opencv學習筆記五十二：基於Haar或LBP級聯分類器的實時人臉人眼檢測

#include<opencv2\opencv.hpp> using namespace cv; using namespace std; int main(int arc, char** argv) { namedWindow("output",

機器學習筆記之十二——SVM原理及推導

svm（support vector machine）是一種二分類演算法，它的目標在於尋找一個能將兩種點分離的直線或平面或超平面。如圖（來自wiki）：圖中的紅線將兩邊資料點分開，這條線就是分割直線，同樣的，在三維座標軸中，將兩邊資料點分開的平面，稱為分割平面；更高維的空間座標軸，

Java Web 學習筆記之十一：RestEasy統一處理異常

JBoss RestEasy框架配置異常統一處理前提利用JBoss restEasy框架搭建的restful java web後臺應用希望通過統一的方式對restful介面丟擲的異常進行

Effective C++讀書筆記之十二：複製物件時勿忘其每一個成分

class Date{...}; class Customer { public: ... private: std::string name; Date lastTransaction; }; class PriorityCustomer:public Customer { public: Pri

《Javascript權威指南》學習筆記之十四：JavaScript內建類

前面的幾篇博文分別介紹了物件、字串、陣列、日期等內建類，本篇將介紹Boolean/Math/Function/Arguments類一、使用Boolean類處理邏輯值 Boolean

Java Web 學習筆記之十四：RestEasy新增Filter過濾器預處理請求

RestEasy新增Filter過濾器預處理請求前提定義filter過濾器，預處理http請求在resteasy框架下配置filter 實現功能攔截http請求，獲取請求頭中的

Maven學習筆記之十九：Maven倉庫(maven倉庫分類)

Maven倉庫(maven倉庫分類) ---------- 對於maven來說，倉庫只分為兩類：本地倉庫和遠端倉庫。當maven根據座標尋找構件的時候，它首先會檢視本地倉庫，如果本地倉庫存在此構件，則直接使用；如果本地倉庫不存在此構件，或者需要檢視是否有更新的構件版本

Androin學習筆記四十二：Java android Socket通訊檢測（server）連線是否斷開

Pre 在利用socket寫通訊程式的時候，想檢測伺服器是否還活著。從網上找了很多資料，都沒有自己合適的，最後自己想了個辦法，不過也相當於截取了心跳檢測的一部分。這裡檢測的是遠端server的連線，而不是本地是否連線成功。首先想到socket類的方法isClosed()、isConnected

CSharp學習筆記之十二委託與事件（續）

尼瑪，當把剛才的文章給儲存發表之後，想看看其他方面你的文章，這時候又看到了另外的一個比較不錯的文章，名字與本題同名，正覺得納悶呢，仔細一看，原來是同一作者的東西，頓時有一種被騙的感覺。這些都不重要，真心的不重要，重要的是這些東西我們都能夠學到比較多的東西。所以，這裡就又將那

Linux運維學習筆記之十七：LNMP的WEB架構深度優化之效能優化

第二十七章LNMP的WEB架構深度優化之效能優化一、Nginx效能優化-配置Nginx的worker程序個數 1、worker的作用和初始設定 worker_processes就是worker角色的程序個數，也就是nginx啟動後有多少個worker處理http請求。ma

kubernetes學習筆記之十二：資源指標API及自定義指標API

第一章、前言

第二章、安裝部署metrics-server

1、下載yaml檔案,並安裝

2、驗證

3.注意事項

第三章、安裝部署prometheus

1.下載yaml檔案及部署前操作

2.安裝node-exporter(用於收集節點的資料指標)

3.安裝prometheus

第四章、部署prometheus+grafana

第五章、實現HPA

1、使用v1版本測試

2、使用v2beat1

3.使用v2beat1測試自定義選項

相關推薦

第五章、實現HPA　　