Understanding alert timers

Hi,

I've been using Solarwinds for some time, but still having trouble understanding "Evaluation Frequency of Alert" and it's relation to "Condition must exist for more than".

I believe if the "Evaluation frequency" is set to 1 minute, and the "Condition must exist for more than" is left to it's default, then condition must exist for 10 minutes.

I'm including our node settings to check if this has any influence on the triggers....

Node settings:

Node Status Polling:  120 seconds  
Collect Statistics Every:  10 minutes

If what I understand is true, then the node will be checked in the database every 1 minute for a the alert condition. If the alert condition is met, Solarwinds will continue to check for 10 minutes and if the condition happens again during that 10 minutes (10 evaluations) the alert will trigger?

Thank you, Pat

Parents
  • I believe you are correct. In your scenario once the condition is met the alert is put into Warning status. It will get polled 10 more times before the alert triggers. If lets say on the third poll the node responds then a Reset is issued. If after the 11th poll it is still down it will escalate the Warning to an Alert.

  • Thanks for the response Bob - now I'm wondering if the alert is triggering right when it gets a response higher than the set threshold. Then, if it happens to not get a response less than the threshold for 10 tries, it will reset the alert.

    -Pat

  • You can look in Events and filter for Alerts Triggered and Alerts Reset to see when they are getting tripped. Another option is to set up a new alert that trips and sends an email whenever the node is in a Warning state and have it also send an email on the reset. We use those type alerts on some of our 'special' nodes (such as out internet circuit) so that we know right away there may be something going on without needing to wait the full X minutes for the actual alert to trip. 

Reply
  • You can look in Events and filter for Alerts Triggered and Alerts Reset to see when they are getting tripped. Another option is to set up a new alert that trips and sends an email whenever the node is in a Warning state and have it also send an email on the reset. We use those type alerts on some of our 'special' nodes (such as out internet circuit) so that we know right away there may be something going on without needing to wait the full X minutes for the actual alert to trip. 

Children
No Data