How to Use kubectl autoscale for Kubernetes Pods

Use kubectl autoscale to create a Horizontal Pod Autoscaler (HPA) that automatically adjusts the number of pods in a deployment based on CPU usage or other metrics. The basic command is kubectl autoscale deployment [name] --min=[minPods] --max=[maxPods] --cpu-percent=[targetCPU]. This helps keep your app responsive and efficient.

Syntax

The kubectl autoscale command creates a Horizontal Pod Autoscaler for a deployment or replicaset. Here is the syntax:

• kubectl autoscale deployment [DEPLOYMENT_NAME]: Targets the deployment to autoscale.
• --min=[MIN_PODS]: Minimum number of pods to keep running.
• --max=[MAX_PODS]: Maximum number of pods allowed.
• --cpu-percent=[TARGET_CPU]: Target average CPU utilization percentage to maintain.

bash

kubectl autoscale deployment [DEPLOYMENT_NAME] --min=[MIN_PODS] --max=[MAX_PODS] --cpu-percent=[TARGET_CPU]

Example

This example creates an autoscaler for a deployment named webapp. It keeps at least 2 pods, scales up to 5 pods, and targets 50% CPU usage per pod.

bash

kubectl autoscale deployment webapp --min=2 --max=5 --cpu-percent=50

Output

horizontalpodautoscaler.autoscaling/webapp autoscaled

Common Pitfalls

Common mistakes when using kubectl autoscale include:

• Not having resource requests set on pods, so CPU usage can't be measured properly.
• Setting --min higher than --max, which causes errors.
• Using kubectl autoscale on unsupported resources like pods directly instead of deployments or replicasets.
• Expecting immediate scaling; autoscaling reacts based on metrics over time.

bash

kubectl autoscale deployment webapp --min=5 --max=3 --cpu-percent=50
# This will fail because min is greater than max

# Correct usage:
kubectl autoscale deployment webapp --min=2 --max=5 --cpu-percent=50

Output

error: min cannot be greater than max
horizontalpodautoscaler.autoscaling/webapp autoscaled

Quick Reference

Option	Description
--min	Minimum number of pods to run
--max	Maximum number of pods allowed
--cpu-percent	Target average CPU utilization percentage
deployment [NAME]	Name of the deployment to autoscale

Key Takeaways

Use kubectl autoscale to create a Horizontal Pod Autoscaler for deployments.

Always set resource requests on pods for accurate CPU-based autoscaling.

Ensure the minimum pod count is less than or equal to the maximum pod count.

Autoscaling adjusts pods based on CPU usage over time, not instantly.

kubectl autoscale works on deployments or replicasets, not individual pods.