In a situation where the Virtual IPS Sensor is not working, the traffic is load balanced by the other active Sensors. The public cloud auto scaling option also provides for Sensor failover in which a new Sensor instance is launched to load balance the traffic when one Sensor fails.
The Virtual Probe pings the Sensor multiple times within a second to check the status of the Sensor. In a scenario when the Sensor is not reachable, the Probe immediately directs the packet to another Sensor. This allows for rapid failure detection and rerouting of traffic to the next available Sensor. When there is no Sensor available, due to the fail-open functionality in the Probe, there is no inspection of packets.