This repository contains a kubernetes aware monitor that measures http keep-alive idle timeout settings of ingress backend services. The measurements are exported as prometheus metrics.
This monitor can be used to ensure backends have a http keep-alive idle timeout configured that is as least as high as the ingress reverse proxy uses for the upstream connections.