I believe that internally we have a snapshot query that runs every 10 or 15 minutes against the

system_info

table to get host_uuid and we use that as a heartbeat