Routers and Health Checks?

This is more of a RFE/question.

Right now it looks like neither the HAProxy or F5 router plugins take advantage of health checks to stop routing traffic to specific pods. Openshift is polling these via liveness probes (at what frequency?), and will eventually kill a pod, which triggers it being removed from the service, which will then be detected by the router plugin and synced with the router configuration.

Is the latency of this process already low enough (<5s) that additional health checks would not be valuable in the router?

Both HAProxy and F5 have internal support for a variety of health checks that can be used to add/remove backends from a loadbalanced group. And we are configuring liveness probes for our pods. Should liveness probes be added to the router as health checks when a route is created?