Our load balancer is set up with health checks against all nodes in our Hybris Commerce cluster. The checks are done very regularly and creates a lot of sessions, which eventually time out. But the servers have to maintain a fairly large number of sessions, cause by the health checks.
Is it recommended to do this in a different way? One could of course make the health check work on the TCP layer, but that reduces the confidence of the result. Is there a URL or argument suitable for this?