We're experimenting with Cacti and so far we are pleased with it's
monitoring capabilities. However, we've seen a handful of our NW6.5 sp5
servers experience extremely high CPU utilization after polling SNMP with
Cacti. The servers that experience the high CPU utilization become
essentially unusable until we take the server out of Cacti's polling
queue. A few minutes later the CPU utilization returns to normal. I came
across an older thread in the forums describing this phenomenon


but unfortunately that thread never seemed to have found a solution. Does
anyone else know about this issue or have an ideas that might help us
resolve this issue? Thanks in advance for your help.