[mike_d]

So they lost connectivity but were not rebooted (as in your uptime didn't go down)?

Sounds like they had enough battery to safely pause all the instances and then resume them when power came back.