Explanation

To meet the requirements of scaling Amazon EKS according to workload with the least operational overhead, the correct combination is:

B. Use the Kubernetes Metrics Server to activate horizontal pod autoscaling. C. Use the Kubernetes Cluster Autoscaler to manage the number of nodes in the cluster.

Why this combination works:

1. Horizontal Pod Autoscaling (HPA) with Kubernetes Metrics Server:

   • HPA automatically scales the number of pods in a deployment based on observed CPU utilization or custom metrics
   • Kubernetes Metrics Server collects resource metrics from Kubelets and exposes them via the Metrics API
   • This handles scaling at the application/pod level based on actual workload

2. Kubernetes Cluster Autoscaler:

   • Automatically adjusts the size of the EKS cluster by adding or removing nodes
   • Works in conjunction with HPA - when pods can't be scheduled due to insufficient resources, Cluster Autoscaler adds nodes
   • When nodes are underutilized, it removes nodes to save costs
   • This handles scaling at the infrastructure/node level

Why other options are incorrect:

A. Use an AWS Lambda function to resize the EKS cluster:

• This would require custom code, monitoring, and maintenance
• Higher operational overhead compared to native Kubernetes solutions
• Not the "least operational overhead" approach

D. Use Amazon API Gateway and connect it to Amazon EKS:

• API Gateway is for creating, publishing, and managing APIs
• Not related to autoscaling EKS clusters
• This would not help with scaling based on workload

E. Use AWS App Mesh to observe network activity:

• App Mesh is a service mesh for monitoring and controlling microservices
• Provides observability for network traffic but doesn't handle autoscaling
• Not designed for automatic scaling of EKS clusters

Key Benefits of the Correct Solution:

• Native Kubernetes integration - uses standard Kubernetes components
• Automated scaling - both at pod and node levels without manual intervention
• Cost optimization - scales down during low usage periods
• Minimal operational overhead - managed by Kubernetes and AWS
• Seamless integration - HPA and Cluster Autoscaler work together automatically

This combination provides a complete autoscaling solution where HPA scales pods based on workload, and Cluster Autoscaler adjusts the node count to accommodate the pods, all with minimal operational effort.