Question #1541
A company uses Amazon Elastic Kubernetes Service (EKS) to deploy containerized applications. The workload experiences significant fluctuations, and the company needs the cluster to automatically scale resources with minimal manual intervention. Which two steps should be taken to achieve this goal? (Choose two.)
Integrate Amazon Route 53 to distribute traffic and trigger scaling events.
Implement Kubernetes Horizontal Pod Autoscaler (HPA) using the Kubernetes Metrics Server.
Configure an AWS Auto Scaling group to directly manage the number of nodes in the EKS cluster.
Deploy the Kubernetes Cluster Autoscaler to dynamically adjust the node count.
Use Amazon CloudWatch alarms to trigger AWS Lambda functions for scaling nodes.
Explanation
To achieve automatic scaling in EKS:
- A (HPA with Metrics Server): Scales pods horizontally based on CPU/memory usage. The Metrics Server provides necessary metrics for HPA to make scaling decisions.
- C (Cluster Autoscaler): Adjusts the number of nodes in the cluster when pods cannot be scheduled due to resource constraints, working with ASGs.
Why others are incorrect:
- B: ASGs alone cannot dynamically scale nodes without Cluster Autoscaler.
- D: Using Lambda with CloudWatch is less efficient than Cluster Autoscaler.
- E: Route 53 handles DNS, not scaling.
Key Points: Use HPA for pod scaling and Cluster Autoscaler for node scaling to handle fluctuating workloads in EKS.
Answer
The correct answer is: AC