We're experiencing chronic intermittent alert flapping issues with SAM snmp-based process monitors on Linux managed nodes.
The problem will occur and the monitor will flap back and forth between up and down status for a variable amount of time (30 min - an hour) and then suddenly stop.
We have adjusted polling intervals, examined poller, primary, database and managed node resource utilization and set delays in alert action processing, but it still causes false alerts on a regular basis.
The problem primarily appears on RedHat Enterprise Linux servers of different OS and patch levels running the Net-SNMP agent (also at levels up to 5.7.2); especially on Oracle servers whree we're monitoring multiple iterations of the listener process and ora processes.
Perl-based script monitors setup on the same targets to monitor these same processes do not alert (since the processes are not really going down and the script monitor successfully retrieves & parses the PS table).
Debug does not seem to log the actual result of the SNMP Get_Table query, which would be nice to trap an examine for truncation or corruption of some sort.
Does anyone know how to enable logging either in the monitor or elsewhere in the product to record the SNMP process monitor query results?