This discussion has been locked. The information referenced herein may be inaccurate due to age, software updates, or external references.

You can no longer post new replies to this discussion. If you have a similar question you can start a new discussion in this forum.

Hundreds of HW sensor alerts, all at the same time and the network object is the sensor.

yaquaholic over 10 years ago

We had a flood of hardware sensor alerts last night at 17:03, 546 of them, but each alert didn't detail the network device name, but instead the hardware sensor name.

Now two things are worrying me, firstly why the alerts are coming through without Node Names, as the network object names are listed as the hardware component names (see below):

The Advanced Alert is set up to record to NetPerfMon Event Log, with: ${SensorDisplayName} on ${NodeName} is ${SensorOrionStatusDescription}

When I look in the Event Log they appear, but why aren't they formatted with their node names in the Alert log

Is this normal behaviour?

Secondly is why we received 546 alerts, with the same time stamp (23/09/13 17:03) from our ESX hosts, which apart from these alerts appear to be happy.

Any ideas?

Top Replies

0 njoylif over 10 years ago

The alerts page out of the box leaves something to be desired, IMHO...
so I spent some time writing this page up - which will relate back to the node the hardware belongs to.
Just create a view with a report as the page, put it on your appropriate tabs and there you have it.
As to why you have all the alerts all of the sudden, I have no idea. try doing a SNMPWalk on the systems alerting and see what is coming back. Then you'll have a place to start troubleshooting.
Cancel
Vote Up +2 Vote Down

Cancel
0 yaquaholic over 10 years ago in reply to njoylif

Thanks for the input, and your solutions looks perfect for us.
But I'm getting issues, namely:
Msg 207, Level 16, State 1, Procedure RH_HCOpsAlertsNAck, Line 120
Invalid column name 'hname'.
And..
Msg 207, Level 16, State 1, Procedure RH_HCOpsAlerts, Line 120
Invalid column name 'hname'.
In both stored procedures hname is being called as n.hname, where n relates to Nodes.
But there is no hname in the table.
Any ideas what am I doing wrong?
Cancel
Vote Up 0 Vote Down

Cancel
0 njoylif over 10 years ago in reply to yaquaholic

you didn't do anything wrong. my bad. that is a custom field I use that "chops" the caption to just the hostname; this allows reports, etc... to look a little cleaner for us.
replace all instances of "hname" with "caption" and you will be good to go!
Cancel
Vote Up 0 Vote Down

Cancel
0 njoylif over 10 years ago in reply to njoylif

I updated the content share to use caption. Thanks for letting me know!
Cancel
Vote Up +1 Vote Down

Cancel
0 ecklerwr1 over 10 years ago

I too have noticed this behavior and you're correct... without node information these alerts don't mean much. I like the hardware monitoring a lot but I've noticed some of my Cisco devices including a couple nexus switches with TON'S of sensors seem to have the habit of showing the green hardware sensor up messages for every sensor almost every polling cycle. For these select noisy devices I've just disabled hardware monitoring.
I'm curious if anyone else has seen this on devices?
For a loaded Nexus 7k one set of sensor up messages is like 90 events.
Cancel
Vote Up 0 Vote Down

Cancel
0 yaquaholic over 10 years ago in reply to njoylif

Fantastic, thanks
I also noticed there was a ref to i.IntId, which should have been i.InterfaceID, once correct the tables are populated.
Cancel
Vote Up 0 Vote Down

Cancel
0 njoylif over 10 years ago in reply to yaquaholic

ah yes, that is like the hname for us. glad you got it working. I'll tweak that as well.
Cancel
Vote Up +1 Vote Down

Cancel