We are started to use https://github.com/haproxytech/kubernetes-ingress together with a base image of haproxy docker.io/haproxytech/kubernetes-ingress:3.2.4 in AWS EKS Kubernetes cluster.

Configuration - Deployment, 2 replicas. Right now have just 1 HAProxy Ingress with < 10 requests/min and > 150 Nginx Ingresses (which are scanning by HAProxy Ingress controller every sync-period interval):

DEBUG controller/controller.go:364 [transactionID=xxx] ingress 'nginx-ingress-ns/nginx-ingress-name' ignored: no matching ....

From Grafana Dashboard of HAProxy Ingress Controller we are observing that pods memory usage increases linearly (about +8Mb/hour)

Mentioned issue is not observed on another cluster with just < 15 Nginx Ingresses