Time out traceroutes by IP hop, not by full traceroute length

[apike]

Typically, when you do a traceroute, it gives a certain amount of time per IP hop, and if a given IP takes too long it will get kICMPTimeExceeded and continue on. Sometimes though, currently, a hop will fail but neither will the timeout, and it will wait forever. Apparently the timeout and traceroute code is complicated, so for 2.0 Jot is going to do a temporary fix of timing out if the overall traceroute after (say) 30 seconds, in #462.

In a future release, we want to understand how the timeout is supposed to work, and fix it so we time out if an IP hop hasn't happened in (say) 5 seconds, as opposed to potentially having a 0-hop traceroute take 30 seconds to time out, or having a long but progressing traceroute time out because it hit 30 seconds.