Skip to main content
Gå til innhold

HorizontalPodAutoscaler

A HorizontalPodAutoscaler (HPA for short) automatically updates a workload resource (such as a Deployment or StatefulSet), with the aim of automatically scaling the workload to match demand.

The Platon PaaS cluster is running the Metrics Server which allows for CPU/Memory based horizontal autoscaling.

Example​

In this example (based on the Kubernetes HPA docs) the HPA controller will increase and decrease the number of replicas (by updating the Deployment) to maintain an average CPU utilization across all Pods of 50%:

---
apiVersion: autoscaling/v2
kind: HorizontalPodAutoscaler
metadata:
  name: hpa-test-deployment
  namespace: platon-hpa-test-deployment
spec:
  scaleTargetRef:
    apiVersion: apps/v1
    kind: Deployment
    name: hpa-test-deployment
  minReplicas: 1
  maxReplicas: 10
  metrics:
    - type: Resource
      resource:
        name: cpu
        target:
          type: Utilization
          averageUtilization: 50
$ kubectl -n platon-hpa-test-deployment get hpa
NAME                  REFERENCE                      TARGETS  MINPODS MAXPODS REPLICAS   AGE
hpa-test-deployment   Deployment/hpa-test-deployment 0%/50%    1       10      0          6s

Limitations & considerations​

  • HPA’s does not work with DaemonSets
  • HPA scales on CPU used as a percentage of the CPU request, so it only works on pods that set one — see Setting resources