How to implement automatic scaling of pods on a physical server using a horizontal autoscaler?

To implement automatic horizontal scaling of pods on a physical server using Horizontal Pod Autoscaler (HPA) in Kubernetes, you need to:

1. Configure metrics: HPA usually scales pods based on metrics such as CPU load or custom metrics. Ensure that a Metrics Server or another metrics source is running in the cluster.

2. Define resource limits and requests for containers in pods so that HPA can accurately assess the load.

3. Create an HPA object, specifying the minimum and maximum number of pods, as well as the target metric value (e.g., CPU utilization at 50%).

4. Ensure the physical server has sufficient resources for scaling. If resources are lacking, scaling will not occur.

Example HPA manifest for a deployment:

apiVersion: autoscaling/v2
kind: HorizontalPodAutoscaler
metadata:
  name: example-hpa
spec:
  scaleTargetRef:
    apiVersion: apps/v1
    kind: Deployment
    name: example-deployment
  minReplicas: 1
  maxReplicas: 5
  metrics:
  - type: Resource
    resource:
      name: cpu
      target:
        type: Utilization
        averageUtilization: 50

Thus, HPA will automatically increase or decrease the number of pods based on CPU load, ensuring horizontal scalability on the physical server.