cancel
Showing results for 
Search instead for 
Did you mean: 
Create Post
Level 7

Alert Suppression When Pollers are Down

Jump to solution

I have been tasked with figuring out how to get alerts suppressed if/when our pollers go down. Not sure if there is an easy way to do this for the all of the alerts or on the pollers themselves. I know you can place everything into maintenance mode but I think we are looking at if our entire environment goes down including pollers, we do not want alerts to be sent out since there wouldn't be a way to get into the system at that point in order to suppress the alerts. Hope this question makes sense as I am fairly new to using solarwinds. Thanks in advance for any insight.

0 Kudos
1 Solution
The alerting only comes from the primary polling engine. APE's don't trigger alerts and when they go down generally their nodes just stop getting updates, they wouldn't show every node as going down or anything like that, they just get stuck with whatever their last updated data was until the poller comes back.

Normally the only alerts that would be triggering after that point might be alerts where the condition was already recorded as being true but the required time had not passed, like where cpuload is high for at least 30 minutes. Potentially also if you have custom alert logic that references the current time, such a a custom sql logic that triggers when the difference between the lastupdate and the current time goes above X minutes.

It is pretty hard to find yourself in a situation where the alerting engine is actually up and running but everyone is locked out of the web console, and if that were the case you'd just need to kill the alert service on the main poller.

- Marc Netterfield, Github

View solution in original post

1 Reply
The alerting only comes from the primary polling engine. APE's don't trigger alerts and when they go down generally their nodes just stop getting updates, they wouldn't show every node as going down or anything like that, they just get stuck with whatever their last updated data was until the poller comes back.

Normally the only alerts that would be triggering after that point might be alerts where the condition was already recorded as being true but the required time had not passed, like where cpuload is high for at least 30 minutes. Potentially also if you have custom alert logic that references the current time, such a a custom sql logic that triggers when the difference between the lastupdate and the current time goes above X minutes.

It is pretty hard to find yourself in a situation where the alerting engine is actually up and running but everyone is locked out of the web console, and if that were the case you'd just need to kill the alert service on the main poller.

- Marc Netterfield, Github

View solution in original post