Skip to content

Horizontal Pod Autoscaler

In one line

The HPA is a Kubernetes resource that automatically scales the number of Pods in a Deployment based on CPU, memory, or custom metrics.

In simple words

The Horizontal Pod Autoscaler (HPA) is Kubernetes' built-in auto scaler for Pods. "Horizontal" means it scales by adding or removing Pod instances, not by changing the size of existing Pods.

You define an HPA with a target — e.g., "keep CPU usage at 70%". The HPA continuously checks the actual CPU usage. If it is above 70%, it adds more Pods. If it is below 70%, it removes Pods (down to your configured minimum).

The HPA can scale based on CPU, memory, or custom metrics (like requests per second, queue depth). Custom metrics require a custom metrics adapter. The HPA checks metrics every 30 seconds by default and can scale from a minimum to a maximum number of replicas.

Example

bash
# Create an HPA that scales between 2 and 10 Pods
kubectl autoscale deployment my-app --min=2 --max=10 --cpu-percent=70

# Check HPA status
kubectl get hpa my-app

This command creates an HPA for the my-app Deployment, keeping CPU at 70% with 2-10 replicas.

How it works

  1. The HPA controller reads metrics from the metrics API every 30 seconds.
  2. It calculates the desired replica count based on the target metric.
  3. It updates the Deployment's replica count.
  4. The Deployment creates or removes Pods to match the new count.

Common confusions

  • Confused with: Vertical Pod Autoscaler

    The difference: HPA adds or removes Pods (horizontal scaling); VPA increases or decreases CPU/memory per Pod (vertical scaling). You cannot use both on the same resource for the same metric.

Related terms

Decode error messages

Plain-English explanations of common errors — what they mean, why they happen, and how to fix them.

View errors

← Back to glossary