Maintenance alarms don't show up

[d.pereyra]

Hello folks!

 

In recent days I have noticed a strange behavior in one of the spectro server I manage, which also functions as MLS.  I put a device into maintenance modo and alarm don't show up, but the events are generate normaly. For example a pingable with a critical alarm "THE DEVICE HAS STOPPED RESPONDING TO POLLS", i set in maintenance and the CRITICAL still there.

 

I wait almost an hour without changes.. so  I restart the SpectroServer to fix it, but it happens all the time with any device in same SpectroServer.

I check settings and filters for alarm events, but these are events that have not changed. Any idea of why could this happend, sometimes with delay or without update the alarms.

Thanks

Diego MP

[ackjo04]

If a model has an existing Critical "THE DEVICE HAS STOPPED RESPONDING TO POLLS" and then is put into Maintenance Mode, the Critical alarm is not cleared. It is still an active alarm on the model but is now considered a secondary alarm to the Maintenance alarm.

 

[d.pereyra]

Thanks Joseph for reply, this option is in "No", i know the alarms are hide when is in maintenance mode, but in my case is in maintenance with condition maintenance, but the alarm no updated, still the critical in alarms tab.

 

Diego

[ackjo04]

Maintenance alarms are not displayed in the Alarms tab by default. You have to tell the Alarms tab to display maint alarms

[d.pereyra]

I select all severitys to show in alarms tab, the is no filter to hide any alarms.

[d.pereyra]

After a Healt report of 24hs I get the Notification in red, it could be the origin of why alarms have delay to show and update in oneclick, not only maintenance alarms.

 

		Average	Peak	%Over Critical Value
LATENCIES	Log	0.0 sec	0.0 sec	0%
	Poll	0.3 sec	46.8 sec	1%
	Timer	0.2 sec	84.2 sec	1%
	Notification	36.7 sec	5,231.0 sec	10%

[d.pereyra]

Partially solved by disabling "Unmanaged Trap Handling", I checked that too many events are being generated through traps, unknown and some device in particular. Between 6 and 9 traps per second, although not a high amount. And this server is the MLS with Trap Director enable.

 

After set in disable the "Unmanaged Trap Handling" option the visualization and notification of alarms back to normaly, definitely is a performance problem that I need to debug.

 

I hope my problems/ post! are useful for others.