A recent incident had a client change the IP address of the only hub within their network, this caused 3 of the 4 robots attached to that hub to lose their stuff. It took a 6 hour call with support on Martin Luther King Jr, Day to discover this was the fault of an IP address change. (and three support engineers).
Love the perseverance of the support team and given the circumstance that 1 robot was fine while the other three were reporting themselves to be in maintenance mode doesn't scream "hub IP change". But the hub was still on the network which means it was still broadcasting the nimbus udp packets which means the robots should have been able to find the nimbus and attach to the hub through its new IP by discovery on the nimbus.
But that didn't happen.
CA could have saved nearly 10 staff hours on their side and 7 staff hours on our side (just for the call, there was another 2 hours before that trying on our own) by ensuring that the robots listen for those IDP packets and use them to attach to the information they contain whenever their primary hub disappears.