After update of s/w versions from 4.* to 5.* (current ver. 18.104.22.168 both for Sensor and Manager) we begin to receive multiple alerts with SNMP traps about:
Approximately about 300-400 events per week.
We thought that possible cause: lack of capacity of network equipment (sensor with manager aren't connect directly). Though, after updating of equipment this problem wasn't resolved - alerts are still received.
Any ideas why is this happening? Can we trust the work of Sensor if Manager can't rich em?
Try hardcoding the duplex and speed on the management interface on the sensor and switch. The sensors can operate without the manager, but it sounds like you are having a layer 1 issue between the management server and the sensor.