threshold deactivation event

[orddie]

Hey All,

Im trying to monitor the GPU temps on a system.  I have two thresholds configured for the system.

the first being if the value is less then or equal to 65, send an alert.
the second being if the value is greater then or equal to 75, send an alert.

I have tested both threshold's manually by moving both 65 and 75 numbers around and asking the system to re-pull the current values. Each time the testing triggers the alert as expected, but does not resolve the alarm.

examples..
Current temp is 74, i set the too low value to 76 and too high value to 78.  I got the too low alert. When i changed the two low number to 65 and left the too high to 78, the alarm did not resolve itself.
Current temp is 74, i set the too low value to 65 and too high value to 74.  I got the too high alert. When i changed the two low number to 65 and to high number to 75, the alarm did not resolve itself.

the return to normal value is event 100002.  I have created an event processing policy to send the all clear email and resolve the alarm.  According to the Event monitor tab, the all clear is never getting hit.

If i manually resolve the alarm..  it does not trigger again until the temps are outside of 65 or 75

[victor]

Hi!

When you change threshold itself, it loses it's state - from the system point of view, it's like you remove threshold with value 74, and add threshold with value 65. Because this threshold is inactive already, you don't get deactivation event.
In general, it is not recommended to change thresholds to test them - instead, you can change DCI value using transformation script and see if thresholds work correctly. In your case, transformation script could be as simple as

return 60;

to test low temp threshold for example.

Best regards,
Victor

[orddie]

i tried that.  it triggered the to low, but when i removed the return 60;  and allowed it too pool again..  It did not trigger the all clear.

[victor]

Can you please send screenshots of your configuration?

Best regards,
Victor

[orddie]

well while driving home..  the GPU temp alerts triggered as they should.  I changed nothing...  only it happened naturally. 

I thank you for your help here...  wish i could explain the root cause though.