Horizontal Pod Autoscaler in Kubernetes - Practice Problems & Coding Challenges

Challenge - 5 Problems

💻 Command Output

intermediate

Output of HPA status command

What is the output of the following command after deploying an HPA targeting a deployment named webapp with min replicas 2 and max replicas 5, currently scaling to 3 replicas based on CPU usage?

Kubernetes

kubectl get hpa webapp

A

NAME    REFERENCE          TARGETS   MINPODS   MAXPODS   REPLICAS   AGE
webapp  Deployment/webapp   60%/50%   2         5         3          10m

B

NAME    REFERENCE          TARGETS   MINPODS   MAXPODS   REPLICAS   AGE
webapp  Deployment/webapp   70%/50%   2         5         4          10m

C

NAME    REFERENCE          TARGETS   MINPODS   MAXPODS   REPLICAS   AGE
webapp  Deployment/webapp   60%/60%   2         5         2          10m

D

NAME    REFERENCE          TARGETS   MINPODS   MAXPODS   REPLICAS   AGE
webapp  Deployment/webapp   50%/60%   2         5         3          10m

💡 Hint

Look carefully at the TARGETS column showing current CPU usage versus target CPU utilization.

The kubectl get hpa command shows the current CPU usage as the first number in TARGETS and the target CPU utilization as the second number. The REPLICAS column shows the current number of pods running.

Option A correctly shows 60% current usage against 50% target, with 3 replicas running.

🧠 Conceptual

intermediate

Understanding HPA scaling behavior

If a Horizontal Pod Autoscaler is configured with minReplicas=1, maxReplicas=4, and target CPU utilization of 50%, what happens when the CPU usage stays at 30% for a long time?

A The HPA will keep the pods at the current number, ignoring CPU usage below target.

B The HPA will scale down the pods to 1, the minimum number of replicas.

C The HPA will scale up the pods to 4, the maximum number of replicas.

D The HPA will delete all pods because CPU usage is low.

💡 Hint

Think about what the minimum replicas setting means for scaling down.

The HPA will reduce the number of pods when CPU usage is below the target, but it will never go below the minimum replicas configured.

Since CPU usage is 30%, below the 50% target, the HPA will scale down to the minimum of 1 pod.

❓ Configuration

advanced

Correct HPA YAML for CPU autoscaling

Which YAML snippet correctly defines a Horizontal Pod Autoscaler for a deployment named api-server with min 3 replicas, max 10 replicas, targeting 75% CPU utilization?

A

apiVersion: autoscaling/v1
kind: HorizontalPodAutoscaler
metadata:
  name: api-server
spec:
  scaleTargetRef:
    apiVersion: apps/v1
    kind: Deployment
    name: api-server
  minReplicas: 3
  maxReplicas: 10
  targetMemoryUtilizationPercentage: 75

B

apiVersion: autoscaling/v1
kind: HorizontalPodAutoscaler
metadata:
  name: api-server
spec:
  scaleTargetRef:
    apiVersion: apps/v1
    kind: Deployment
    name: api-server
  minReplicas: 3
  maxReplicas: 10
  targetCPUUtilizationPercentage: 75

C

apiVersion: autoscaling/v2
kind: HorizontalPodAutoscaler
metadata:
  name: api-server
spec:
  scaleTargetRef:
    apiVersion: apps/v1
    kind: Deployment
    name: api-server
  minReplicas: 3
  maxReplicas: 10
  metrics:
  - type: Resource
    resource:
      name: memory
      target:
        type: Utilization
        averageUtilization: 75

D

apiVersion: autoscaling/v2
kind: HorizontalPodAutoscaler
metadata:
  name: api-server
spec:
  scaleTargetRef:
    apiVersion: apps/v1
    kind: Deployment
    name: api-server
  minReplicas: 3
  maxReplicas: 10
  metrics:
  - type: Resource
    resource:
      name: cpu
      target:
        type: Utilization
        averageUtilization: 75

💡 Hint

The autoscaling/v2 API uses a metrics array with resource type and target details.

Option D uses the autoscaling/v2 API with the correct metrics field specifying CPU utilization target.

Option D uses autoscaling/v1 which supports only CPU target but with a different field name; it is valid but less flexible.

Options C and D incorrectly target memory instead of CPU.

❓ Troubleshoot

advanced

Troubleshooting HPA not scaling up

You created an HPA for a deployment, but the pods never scale beyond 1 even when CPU usage is high. Which is the most likely cause?

A The maxReplicas is set to 1, limiting scaling to a single pod.

B The HPA minReplicas is set to 1, so it cannot scale beyond that.

C The deployment's resource requests for CPU are not set, so HPA cannot calculate usage properly.

D The deployment has too many replicas already running.

💡 Hint

Check the maxReplicas setting to see if it limits scaling.

If maxReplicas is set to 1, the HPA cannot scale beyond one pod regardless of CPU usage.

Resource requests missing can cause inaccurate metrics but usually do not prevent scaling entirely.

✅ Best Practice

expert

Best practice for HPA metrics configuration

Which approach is best practice when configuring Horizontal Pod Autoscaler metrics for a web application that experiences variable CPU and memory usage?

A Avoid using metrics and set fixed replica counts to prevent unpredictable scaling.

B Configure only memory usage metric because CPU usage fluctuates too much to be reliable.

C Configure multiple metrics including CPU utilization and memory usage with appropriate targets to allow balanced scaling.

D Configure only CPU utilization metric because memory usage is less important for scaling decisions.

💡 Hint

Consider how multiple metrics can improve scaling decisions.

Using multiple metrics like CPU and memory allows the HPA to scale pods based on the most relevant resource pressure, improving responsiveness and stability.

Relying on a single metric may cause suboptimal scaling.