Request header read timeout using AWS elastic load balancer on ubuntu

Since starting to use AWS elastic load balancer we have started to get 40,000 or more "Request header read timeout." errors per day. It appears that this is related to mod_reqtimeout (http://httpd.apache.org/docs/trunk/mod/mod_reqtimeout.html), which appears to be designed to defend against slow loris ddos attacks (http://ha.ckers.org/slowloris/). From what I can see on the network traffic between our apache and the load balancer, the connection is being kept alive across requests (which is probably a good thing). On the other hand, I suspect that perhaps keeping the connection alive might be responsible for the error, as the load balancer is perhaps sending nothing for periods of time longer than our timeout window just to keep the connection alive.