We are seeing an unusual occurrence with robots. When a remote site has either a power or network issue, a small number of servers are unable to bind to the primary IP address resulting in an alarm with the loopback address presenting as the source IP. The robot must be manually stopped/started to resolve the alarm.
The option to hardcode the hostname is not an option for the customer and we did not see this prior to 5.0 robots. Also, in 90% of the cases for one customer (1500+ servers), the robot recovers with no issues.
Has anyone experienced this behavior, if yes, how have you addressed it?