We are using AWS' Elastic Kubernetes Service (EKS) with an nginx-ingress and an NLB (network load balancer). When activating Route53 healthchecks, which monitor targets in the EKS-Cluster, the client and target reset count in the monitoring stats of the NLB increase drastically.
We deactivated all related timeouts in the ingress-controller to remove it from the equation (no effect on the reset count). We believe it could have something to do with the Route53 healthchecks. When deactiving them, the client and target reset count immediately move towards zero. Any thoughts?