Which solution will resolve this issue with the LEAST administrative overhead?
Scale out the nodes by tracking the memory usage.
Use the Kubernetes Cluster Autoscaler to manage the number of nodes in the cluster.
Use an AWS Lambda function to resize the EKS cluster automatically.
Use an Amazon EC2 Auto Scaling group to distribute the workload.
Explanations:
Scaling out nodes by tracking memory usage alone does not address the issue of the cluster not automatically scaling. This method requires manual intervention and does not leverage Kubernetes’ built-in scaling capabilities.
The Kubernetes Cluster Autoscaler automatically adjusts the number of nodes in the cluster based on the needs of the pods, ensuring that there are enough resources available without manual management. This solution directly addresses the scaling issue with minimal administrative overhead.
Using an AWS Lambda function to resize the EKS cluster adds unnecessary complexity and management overhead. It would require custom logic to determine when to scale the nodes, which is not as efficient as using the built-in Kubernetes Cluster Autoscaler.
While using an Amazon EC2 Auto Scaling group can help manage instances, it does not directly integrate with Kubernetes to adjust the number of nodes based on pod requirements. This option would still require additional configuration and monitoring to effectively manage the scaling of pods within the EKS environment.
As far as I’m aware, the answer is:
Use the Kubernetes Cluster Autoscaler to manage the number of nodes in the cluster.