I'm now running 2-node cluster of PRTG and setup the notification (when sensor goes down) by email.

When the master go offline, I can see the alert from Enterprise Console, however, I can't get any email notification for the event.

Is it normal or anything I missed in the configuration?

Article Comments

Hello,

just to check, is the failover node generally able to send emails? To check this, maybe shut down the Master Node, go to the Failover-Node and Test a notification that uses Email. If it fails, please check under Logs->System Events->Notifications for error messages to the email delivery.

best regards.


Feb, 2013 - Permalink

I can send the test email notification in Setup -> Accountsettings -> Notification successfully, however, I still can't receive any notification for the Master Node down.


Feb, 2013 - Permalink

The failover node can send the notification email to me when a monitored device goes down, however, at the same time I can't get any message about the Master Node was down.


Feb, 2013 - Permalink

I see, then please add a notification to the "Cluster Health"-Sensor on the "Cluster Probe Device". It should go into error-state as soon as at least one Cluster node is disconnected (and will send an email with an according trigger). It might be necessary though to have the latest version of PRTG for this (13.1.1).


Feb, 2013 - Permalink

I've set it up already. All Health Sensor (Cluster, Core, Probe and System) under Cluster Probe Device have been set a State Trigger (When sensor Unusual for at least 60 seconds perform Email to Admin).

When a node is disconnected, I can see the all 4 healther sensor still in good condition (green tick), and the only change is the "Outbound Cluster Connections" change from 1# to 0#.

Then I add a Threshold Trigger "When Outbound Cluster Connection (#) channel is Equal to 0 for at least 5 seconds perform Email to Admin", and and I can now receive the notification email when a node goes offline. However, I still not able to know which node has problem and I do need to check by try to connect them one by one. Is there any better way to have more detail in the notification?


Feb, 2013 - Permalink

The Cluster-Health-Sensor goes into a Down-State when one node is disconnected. So the Trigger would need to be changed from "When sensor Unusual" to "When sensor is Down". Then the state trigger should work.
To see which node is disconnected, please simply view the "Cluster-Status"-Page, I'm afraid in the moment that's the only but best place to see this.


Feb, 2013 - Permalink