I have 3 EKS fargate clusters, 2 in us-west-2 (oregon) that I have had for a couple years now; and I recently (in the last 3 months) added one in us-north-1 (stockholm).
API endpoint access is set to private. Each cluster VPC is peered to another VPC in the same region as the fargate cluster running an S3 instance. I ssh into this S3 instance to manage the cluster.
In the the 2 oregon deployments, running kubectl commands, e.g. kubectl version reliably return within 1 second or less; and always has; and i get the expected result:
Client Version: v1.28.1
Kustomize Version: v5.0.4-0.20230601165947-6ce0bf390ce3
Server Version: v1.28.2-eks-f8587cb
However, in the the newer eu-north-1 deployment, it can take up to 20 full seconds to return. The performance of the cluster services itself seems fine, but all calls to kubectl in the Stockholm deployment take on average FAR longer to return than the US ones. However, sometimes the response is quick and returns within 1 second.
Is this variability normal? If it is not normal, what sort of configuration error could possibly be causing it?
AWS OFFICIALUpdated 3 months ago
