Forum Discussion

Nick's avatar
8 years ago

Suppress 'Host Status' alerts when Collector is down

I would like to request the ability for child device alert suppression in the event of a single collector going down, (failover not enabled/available).

(I have been advised to raise this here after discussions with LM Support)

So think of this in a Parent/Child architecture –

if a Collector is monitoring 150 devices and the collector goes down, it is by that very fact unable to monitor the 150 devices so, therefore, LM will be unable to see/receive data from those devices.

At the moment this means that we would receive 151 host down alerts.

If we look at this from the Parent = Collector & Child = devices point of view, we have established that the Child devices are reliant on the Parent Collector to send data to LM, so can the ability, in the event of the Parent Collector failing that any Child Device alerting is automatically suspended/suppressed?

There is obviously an option in the Collector management to enable & specify ‘Failover’ – so devices monitored by the collector failover to another specified collector in the event of the primary failing – this means that there is already logic in place to detect and carry out actions in the event of a collector failing. Can this logic be extended to show only the collector down alert & to suppress the (child) device alerts?

 

  • In the event of a collector being determined to be down, delivery of all alerts from devices monitored by that collector will be suppressed. While the devices will show as being in a critical alert condition in the UI, the 'device down' alerts will not be sent out to any recipients (email, SMS, voice, integrations).
    https://www.logicmonitor.com/support/settings/collectors/what-if-my-collector-goes-down/
    The devices will also appear with a 'collector down' marker as per the 'W7 on iMac' device here:

     

  • Hi @Antony Hawkins,
    Yes, you are correct in that the device alerts are suppressed, however, the 'Host Status' alerts are not, you still get the 'Host Status' alert for each device - alerts appear in dashboard & phone notification is received, all be it throttled (x in x minutes). Obviously, do not want to switch 'Host Status' alerting off so would like this suppressing as well.

  • Thanks @Nick
    As per our call, I agree that having the option to suppress in-application Host Status alerts in the event of a down collector, so they don't clutter dashboards and alert lists, is a legitimate and sensible request.
    As also discussed, please let me know of any examples of any host down alerts being escalated out whilst the corresponding collector was down.

  • I’d be interested in this as well, has this been implemented? The only other option I can think of is to not alert on host down datapoint but that would reduce my systems reliability and not give me a whole picture of monitoring system status. We’ve been having issues where collectors stop collecting for periods of time each day, and will get 100+ alerts for every down device a collector has and is increasing alert fatiguez