We have an MPLS with several sites and right now there is no good way to measure site-to-site latency via LM. We don\'t want to have a collector in one site monitoring a device in another site for obvious reasons. The services feature would be perfect, but that\'s only for monitoring public IPs from external sources. We\'ve been advised to script this functionality and we\'ll probably do so, but it\'s annoying that LM can\'t do this. Use case: we have latency sensitive applications running over the MPLS, and also have a 60ms SLA. Right now when we believe we are experiencing latency issues, we have no historical data to send to the provider, so it becomes an argument in which we\'re saying \'\'something doesn\'t look right\'\' and they say \'\'looks ok now.\'\' We also can\'t easily correlate latency data (because there is none) with network traffic to rule out load.