Hi,
One additional bit of info on this, I’ve just noticed that the clickable link I’ve been using to try to weed out the alerting sensors on the summary table for ‘down’ sensors, actually triggers a search for ‘warning’ not just ‘alerted’!
So… I’ve been trying to clear up a list which is filtered on the wrong criteria actually.
Maybe the URL for ‘down’ sensors should be: /health/metric=sensors/event=alert%2Calert/
Instead of: /health/metric=sensors/event=alert%2Cwarning/
This produces a list which doesn’t contain all the warning sensors I’m trying to get rid of and just leaves the genuinely alerted ones. Turns out I’ve been trying to get them off a list that they are (maybe) supposed to show up on?
Although that said, I’m not sure they should really get to the stage of ‘warning’ if they are set to not alert at all, or is that just a matter of opinion :)
Cheers!
From: observium [mailto:observium-bounces@observium.org] On Behalf Of Robert Williams
Sent: 21 December 2015 13:47
To: Observium Network Observation System (observium@observium.org) <observium@observium.org>
Subject: [Observium] How to shush alarms for sensors that lie...
Sent: 21 December 2015 13:47
To: Observium Network Observation System (observium@observium.org) <observium@observium.org>
Subject: [Observium] How to shush alarms for sensors that lie...
Hi,
The recent changes to the homepage layout have placed more attention on the summary statistics (alerted ports, alerted sensors etc.) which previously were not as visible. A positive side-effect of this has been the setting of more ‘appropriate’ manual
thresholds for many of the less important sensors, to tidy them up. It has revealed one issue though, which I’m not sure how to work around.
Take the following example:
As you can see, the (unrealistic) dBm being reported by the Cisco chassis is stuck in a ‘warning’ state, despite having alarms disabled, the threshold on manual and thresholds set well clear of the 8.1dBm which it is reporting.
(This port actually has a copper-converter in it, hence, the dBm is just rubbish spat out by an adapter which doesn’t understand copper)
So despite my best efforts to silence this warning, it still persists as a ‘down’ sensor, on the summary page showing up in red. (Not ignored, but actually down). We have quite a few of these adapters as well as a large number of voltage sensors (again,
courtesy Cisco entity SNMP crapness) which are being reported as zero volts. These are also set to ‘ignore’ but continue to show as ‘down’ on the new summary homepage, some examples below:
So – Does anyone know how I can totally and completely ignore/hide the broken sensors like these?
Cheers!
Robert Williams
Custodian Data Centre
Email: Robert@CustodianDC.com
Robert Williams
Custodian Data Centre
Email: Robert@CustodianDC.com