For many performance based tests (not fault based tests like ping), it would be very helpful to have an HM feature that allows the test not go bad if the test returns a value for x test intervals. As an example, if using an snmp get or perf counter test to check the CPU load of a server, HM goes into alarm on the first occurrence of the CPU being > 90% (as an example). How about having a setting in the test properties dialog box to allow the user to select how many tests have to be performed prior to making the test report ‘bad’ as its status? As an example, if you want to be alarmed only if the server CPU has been above 90% for 5 minutes or 10 minutes (five/ten consecutive tests if the test interval is at 1 minute).
I know you can set the alarms to auto retest upon alarm (using alert profile), but this retests immediately, so will not do. Also, you could reset the test interval, but then you would miss the other tests that should have been checked in this example. So, I would consider these to be more of a workaround/kluge instead of then required feature.
All day long we have various tests that are alarming (average 6 to 10 alarms at any time), these alarms are being ignored by the IT staff since they grow numb of seeing performance based tests going in and out of alarm all day.
I think it would be much more useful to many in IT if they knew that the CPU (or network utilization or pick anything performance related) had been pegged for the last 5 to 10 minutes.
Can I get other users thoughts on this?
Thank you