When the agent wait-for bootstrap-complete command starts, it first tries to establish contact with the assisted-service on the rendezvous host. This might reasonably take some time, since we don't know the mechanism by which the user will attempt to boot the hosts (sneakernet remains a valid option), nor the relative timing of this with the running of the command.

Currently, during this process we output a message every 30s reading:

Waiting for cluster install to initialize. Sleeping for 30 seconds

and during this time we are eating into the overall bootstrap-complete timeout. It would be better if we printed only one message, like

Waiting up to 30m00s for rendezvous host to initialize

and once the host appears, enter a separate timeout phase for the bootstrapping.