AutoScaler - HPA

HPA, VPA, CA
강의자료 PDF
Autoscaler with Horizontal Pod Autoscaler for Kubernetes.


1. Metrics Server 설치

1-1) Metrics Server 다운 및 설치

kubectl apply -f https://raw.githubusercontent.com/k8s-1pro/install/main/ground/k8s-1.27/metrics-server-0.6.3/metrics-server.yaml
kubectl apply -f https://github.com/kubernetes-sigs/metrics-server/releases/download/v0.3.6/components.yaml

Metrics Deployment 수정 
# 추가 작업 없음
# args 에 `kubelet-insecure-tls` 추가
kubectl edit deployment metrics-server -n kube-system

#spec:
#  containers:
#  - args:
#    - --cert-dir=/tmp
#    - --secure-port=443
#    - --kubelet-preferred-address-types=InternalIP,ExternalIP,Hostname
#    - --kubelet-use-node-status-port
#    - --metric-resolution=15s
#    - --kubelet-insecure-tls  <-- 추가
# args 에 `kubelet-insecure-tls` 와 `kubelet-preferred-address-types=InternalIP` 추가
kubectl edit deployment metrics-server -n kube-system

#spec:
#   containers:
#   - args:
#     - --cert-dir=/tmp
#     - --secure-port=4443
#     - --kubelet-insecure-tls                       <-- 추가
#     - --kubelet-preferred-address-types=InternalIP <-- 추가

설치 확인 (True값 확인)

kubectl get apiservices |egrep metrics

------------------------
v1beta1.metrics.k8s.io                     kube-system/metrics-server   True        28m
------------------------

메트릭 값 확인 (1~2분 후)

kubectl top node

------------------------
NAME             CPU(cores)   CPU%   MEMORY(bytes)   MEMORY%   
k8s-master       485m         9%     4852Mi          32%       
k8s-node1        413m         8%     4929Mi          33%       
k8s-node2        554m         11%    4672Mi          31%       
------------------------


2. HPA (Horizontal Pod Autoscaler)

Autoscaler with Horizontal Pod Autoscaler CPU for Kubernetes.

2-1) Target Deployment (CPU) / Service

apiVersion: apps/v1
kind: Deployment
metadata:
 name: stateless-cpu1
spec:
 selector:
   matchLabels:
      resource: cpu
 replicas: 2
 template:
   metadata:
     labels:
       resource: cpu
   spec:
     containers:
     - name: container
       image: kubetm/app:v1
       resources:
         requests:
           cpu: 10m
         limits:
           cpu: 20m
---
apiVersion: v1
kind: Service
metadata:
 name: stateless-svc1
spec:
 selector:
    resource: cpu
 ports:
   - port: 8080
     targetPort: 8080
     nodePort: 30001
 type: NodePort

2-2) HPA - Resource (Utilization)

apiVersion: autoscaling/v2
kind: HorizontalPodAutoscaler
metadata:
  name: hpa-resource-cpu
spec:
  maxReplicas: 10
  minReplicas: 2
  scaleTargetRef:
    apiVersion: apps/v1
    kind: Deployment
    name: stateless-cpu1
  metrics:
  - type: Resource 
    resource:
      name: cpu
      target:
        type: Utilization
        averageUtilization: 50

kubectl get hpa -w

while true;do curl 192.168.56.30:30001/hostname; sleep 0.01; done
while true;do curl 192.168.0.30:30001/hostname; sleep 0.01; done
kubectl delete horizontalpodautoscalers.autoscaling hpa-resource-cpu


Autoscaler with Horizontal Pod Autoscaler Memory for Kubernetes.

2-3) Target Deployment (Memory) / Service

apiVersion: apps/v1
kind: Deployment
metadata:
 name: stateless-memory1
spec:
 selector:
   matchLabels:
      resource: memory
 replicas: 2
 template:
   metadata:
     labels:
       resource: memory
   spec:
     containers:
     - name: container
       image: kubetm/app:v1
       resources:
         requests:
           memory: 10Mi
         limits:
           memory: 20Mi
---
apiVersion: v1
kind: Service
metadata:
 name: stateless-svc2
spec:
 selector:
    resource: memory
 ports:
   - port: 8080
     targetPort: 8080
     nodePort: 30002
 type: NodePort

2-4) HPA - Resource (AverageValue)

apiVersion: autoscaling/v2
kind: HorizontalPodAutoscaler
metadata:
  name: hpa-resource-memory
spec:
  maxReplicas: 10
  minReplicas: 2
  scaleTargetRef:
    apiVersion: apps/v1
    kind: Deployment
    name: stateless-memory1
  metrics:
  - type: Resource 
    resource:
      name: memory
      target:
        type: AverageValue
        averageValue: 5Mi

while true;do curl 192.168.56.30:30002/hostname; sleep 0.01; done
while true;do curl 192.168.0.30:30002/hostname; sleep 0.01; done
kubectl delete horizontalpodautoscalers.autoscaling hpa-resource-memory


Sample yaml

* HPA for - Pods

apiVersion: autoscaling/v2
kind: HorizontalPodAutoscaler
metadata:
  name: hpa-pods
spec:
  maxReplicas: 10
  minReplicas: 2
  scaleTargetRef:
    apiVersion: apps/v1
    kind: Deployment
    name: stateless-app1
  metrics:
  - type: Pods
    pods:
      metric:
        name: packets-per-second
      target:
        type: AverageValue
        averageValue: 10

HPA - Object

apiVersion: autoscaling/v2
kind: HorizontalPodAutoscaler
metadata:
  name: hpa-object
spec:
  maxReplicas: 10
  minReplicas: 2
  scaleTargetRef:
    apiVersion: apps/v1
    kind: Deployment
    name: stateless-app1
  metrics:
  - type: Object
    object:
      metric:
        name: requests-per-second
      target:
        type: Value
        value: 10
      describedObject:
        apiVersion: networking.k8s.io/v1beta1
        kind: Ingress
        name: ingress-hpa


Referenece

Kubernetes

Others