I am working on migrating a Docker Spring Boot Reactive application from AWS ECS to AWS EKS.

When running the application in AWS EKS, I notice that all AWS SDK calls (DynamoDB Async updates, Firehose puts, etc) are taking 5-6x latency in Kubernetes EKS, compared to ECS.

I have ensured that the hardware configuration (EC2 host) is the same. I also confirmed that the available memory for the pod in EKS is same as the task in ECS.

How can I debug this? Is there a difference in the way http calls are made from Docker container in EKS and Docker container in ECS?

AWS SDK version used: 2.17.220