Hello,
We are monitoring many Windows servers using SNMP. During parts of the night these virtual services get backed up and will have a snapshot taken and removed. During this process the server can freeze for 20 seconds which is ok, but LibreNMS will send and alert it’s down then recovered.
How can I make this less harsh and not get an alert if it can’t poll for 10-20 seconds? If it’s 45 and above then yes we do want an alert.
Thanks
Hi @gonzo
Just add a delay to your alert rule (at least 1 poll). So if you have 5 minute polling, add 6 minutes delay for instance. Works better with shorter (1 or 2 minutes) polling interval.
Other option is to tweak the SNMP timeout and repeat interval to make it tolerant to 30 seconds freeze. Like 15 seconds timeout and 4 repeat). You’ll probably need to do the same on ping timeout (cause the freeze can occur during both).
This is of course incompatible with a 1 minute polling interval but can be done with a 5 minute polling interval.
Mine all seem to be set to:
Max Alerts = 1
Delay = 1m
Intervals = 5m
So I should set to this below?
Max Alerts = 1
Delay = 6m
Intervals = 5m
Hi @gonzo
It depends on your polling interval and duration.
If you have a 5 minute polling interval and if your polling duration is fast enough (let say below 1 minute), then you should be OK with this :
Max Alerts = 1
Delay = 6m
Intervals = (doesn't matter when MaxAlerts == 1)
If it is still not OK, because your polling duration is between 1 minute and 4`59 then this should do the trick :
Max Alerts = 1
Delay = 10m
Intervals = (doesn't matter when MaxAlerts == 1)
Thanks, where do I find the polling duration?
This topic was automatically closed 91 days after the last reply. New replies are no longer allowed.
