Thresholds can be set to alert the user by raising a system fault when the set threshold value is crossed. For example, a system fault is raised when the Sensor CPU usage crosses an upper threshold value such as 90%.
You can set thresholds at the Global level of the parent (root) admin domain, at the Global level of a child admin domain, and for an individual device. Use the Thresholds tab to change metrics threshold values and choose whether to raise a system fault when a threshold is crossed. The Manager compares each threshold with the average value over one minute.
.png?sv=2026-02-06&spr=https&st=2026-11-20T12%3A38%3A38Z&se=2026-11-20T12%3A49%3A38Z&sr=c&sp=r&sig=v6A6wJXQG9tmptTOU%2FPWU1mP4OqM5YTE1%2BP%2BxlbSEao%3D)
Thresholds can be set for CPU usage, sensor throughput, port throughput, L2 error drop (Layer 2 here refers to too many CRC and Ethernet Packet Compliance errors) and L3/L4 error drop (Layer 3/4 here refers to too many checksum and protocol errors), and memory usage.
Note
When a device restarts, the Manager clears the threshold faults raised for that device. If a metric is still beyond its threshold after the restart, a new fault is raised.