I'm running an ECS cluster behind a Network Load balancer hosting over 200 websites. I've noticed that over time the response time gets worse ( much slower and timeouts ) and requires me to relaunch the Traefik containers to solve it. Initially, I thought it was something with the sites; however, when I hit the sites directly they are quick to respond. I have Traefik 1.7.18 spread across 5 ECS instances which are part of the cluster. I should note that the site containers are on the same instances as where Traefik is deployed. Has anyone else noticed this issue? Is there a memory leak in this version of Traefik?