HPA YAML

Learn the HPA YAML structure.

HPA YAML

An autoscaling/v2 HPA defines the target object with scaleTargetRef, the bounds with minReplicas/maxReplicas, and the list of criteria with metrics:

apiVersion: autoscaling/v2
kind: HorizontalPodAutoscaler
metadata:
  name: api
spec:
  scaleTargetRef:
    apiVersion: apps/v1
    kind: Deployment
    name: api
  minReplicas: 2
  maxReplicas: 10
  metrics:
    - type: Resource
      resource:
        name: cpu
        target:
          type: Utilization
          averageUtilization: 70

Multiple metrics

An HPA can take several metrics; a separate replica count is calculated for each and the largest is chosen. Memory, custom application metrics (requests/second) and external metrics (queue length) can be used too:

  metrics:
    - type: Resource
      resource:
        name: cpu
        target: {type: Utilization, averageUtilization: 70}
    - type: Resource
      resource:
        name: memory
        target: {type: AverageValue, averageValue: 400Mi}

Exercise Write the HPA YAML

Write an HPA YAML: autoscaling/v2; name api; target apps/v1 Deployment api; minReplicas: 2, maxReplicas: 10; a Utilization target of averageUtilization: 70 for CPU.

Show example solution

apiVersion: autoscaling/v2
kind: HorizontalPodAutoscaler
metadata:
  name: api
spec:
  scaleTargetRef:
    apiVersion: apps/v1
    kind: Deployment
    name: api
  minReplicas: 2
  maxReplicas: 10
  metrics:
    - type: Resource
      resource:
        name: cpu
        target:
          type: Utilization
          averageUtilization: 70

Test yourself

When using an HPA, what is recommended for the replicas field in the Deployment YAML?
- Removing it; otherwise each apply overwrites the HPA's value
- Writing a high value
- Writing 0
- Writing the same as maxReplicas
Show answer

Removing it; otherwise each apply overwrites the HPA's value — The HPA now owns the desired replica count.

Specify the object to scale.

spec:
  ____:
    apiVersion: apps/v1
    kind: Deployment
    name: api

Show answer

spec:
  scaleTargetRef:
    apiVersion: apps/v1
    kind: Deployment
    name: api

Match.

autoscaling/v2

HPA API version
scaleTargetRef

Target Deployment
metrics

Metric list
multiple metrics

The largest result is chosen

Next lesson: CPU threshold

We Are Enhancing Your Experience

Cookie Preferences

Strictly Necessary Cookies

Analytical Cookies