feat(uptime): monitor tree

env: server with some services

now:
server(ICMP) per 20s
serviceA(TCP) per 1min
serviceB(HTTP) per 1min

new:
server(ICMP) per 1min
├serviceA(TCP) per 1min
└serviceB(HTTP) per 1min

if A and  B are down, server monitor change to per 20s.
if server is down, stop monitor A&B.

 With logic like this, optimize the monitoring of stress and the number of notifications.
