Updated March 11th, 2016
SNMP Process monitors sometimes cause false alerts (process reported as Down when it is not).
All SAM versions
When the process coincidentally starts or stops in conjuction with the poll, the device does not send the complete process table and SAM reports some processes as Down. This is typically fixed on the next
Workaround: Change the polling method in the Configuration file to either BulkColumn or None. This may fix this issue. Be aware that this change has a global effect on all SNMP process monitors and can decrease performance on increased network traffic.
Additional Information: PollMethod – Specifies the method to use for reading data from SNMP. The default value None uses getNext in the SNMP request. In some environments, it may be beneficial to set this value to either BulkRow or BulkColumn. In either case, SNMP requests will be using getBulk and this should improve performance. The difference between using BulkRow versus BulkColumn is in the way data is read. That is, whether you are reading SNMP tables by rows or by columns.