Trigger Flapping

[mattsmith]

Hi All,

I've read various posts and documentation on trying avoid the change of trigger state each upon each change in data.

My reason for this came when we had a CPU Load problem, the servers load changing between 9.50 and 11, this was happening very frequently and created several alerts as my threshold was >10

I have since changed the trigger to:
({TRIGGER.VALUE}=0&{myserver:system.cpu.load[,avg1].min(#6)}>10) | ({TRIGGER.VALUE}=1&{myserver:system.cpu.load[,avg1].min(#120)}>10)

This should hopefully mean that I will see the first problem after 6 values @ >10 and then this state wont change back to OK until 120 values have be less than 10.

Is this correct? or is there a better way of trying to achieve less events being generated?

Please note that I am using the minimum value with the count option.

many thanks
Matt

[mattsmith]

Hi All,

Sorry to bump this one, has anyone else out there had any scenarios when flapping detection would have been useful? and if so how did people resolve this?

Using averages, trigger hysteresis, etc can help with this but none truly solve the problem. We are still looking for the perfect solution :-)

Any Ideas?

Matt