Kubernetes:HorizontalPodAutoscaler

lastmod 2024-06-11created 2024-06-112 edits

Kubernetes also has a built-in basic autoscaling feature called HPA (Horizontal Pod Autoscaler). HPA increases or decreases the number of pods run by a Deployment based on CPU utilization.

Example

kubectl apply -f autosclinging.yaml

yaml file:

apiVersion: autoscaling/v1
kind: HorizontalPodAutoscaler
metadata:
  name: kubernetes-simple-app-hpa
  namespace: default
spec:
  maxReplicas: 10
  minReplicas: 1
  scaleTargetRef:
    apiVersion: extensions/v1beta1
    kind: Deployment
    name: kubernetes-simple-app
  targetCPUUtilizationPercentage: 30 

A rough calculation of how autoscaling is performed is as follows.

TargetNumOfPods = ceil(sum(CurrentPodsCPUUtilization) / Target)

See also

• k8s

Favorite site

• Part 9!! Kubernetes autoscaling