* Switch wedge agent from reporting CN UUIDs to CN hostnames, and use CheckMK's piggyback mechanism to send wedge status to the correct Host in CheckMK.
* Improve how local ports are selected during connect() attempts, so there's (much) less likely to be conflicts on subsequent runs, due to length of ephemeral port expiry.
* Increase connect() time from 100ms to 1000ms, to deal better with potentially slow/overloaded VMs.
