Fix L2 request generation and improve rate adaptation for HLT requester