cancel
Showing results for 
Search instead for 
Did you mean: 
Create Post
Level 9

Issue with APC upsAdvBatteryCapacity value

Jump to solution

I have configured a UnDP to retrieve the value of upsAdvBatteryCapacity from our APC units. When I test the MIB value from the UnDP utility, I get the expected value of 100 from each of 3 units (meaning 100% capacity remaining). When I configured the alert for this poller to trigger when battery capacity is at or below 40, I received the alert unexpectedly for two of the three units. I populated the message body with the variables ${CustomPollerStatus.Status}, ${CustomPollerStatus.Rate} and ${CustomPollerStatus.Raw} just to see what was coming back. For the unit that did not generate an alert, "raw" and "status" both show 100 and "rate" is empty.

For the two that did trigger, all three variables display the same value, 29 for the first unit and 18 for the second. If the MIB value is returning 100, why are the variables returning a different value? What am I missing here?

Jim

Labels (1)
0 Kudos
1 Solution

Sorry deverts, meant to reply weeks ago. As blsanner mentioned, these gauges are built-in. Just add the desired gauge as he indicated and then edit the resource, and pick the appropriate UnDP that you created to associate to the gauge. If you need to reverse the gauge, for example, in the case of battery level you want 10% to be critical rather than 90%, click the "Show" button beside the Advanced section header (on the resource config) and then select "Reverse Thresholds".

pastedImage_0.png

View solution in original post

0 Kudos
9 Replies
Level 9

I now know where the odd values are coming from....29 is the battery temperature on one and 18 is the load on the other unit. So my trigger conditions are:

Poller Name is equal to upsAdvBatteryCapacity

AND

Status is less than or equal to 40

But the alert is grabbing random UnDP values, not just upsAdvBatteryCapacity as desired. How to I confine the "status" line to just upsAdvBatteryCapacity?

0 Kudos

check your UnDP -- you should be going a GET of the specific MIB variable you need; it defaults to getNext, so if the variable disappears something random will be returned and stored.

Richard,

I checked the query and you are correct it was at the default of "GetNext". I did change it to see if it had any effect, but the issue still remains. The strange thing is that the values are displayed properly on the device pages. I added a "gauge" resource to each device view and assigned the corresponding UnDP. So all gauges display the appropriate value for the UnDP referenced (all capacities are 100%) on each unit. Only the alert itself seems to want to grab a value for a different UnDP. And it is consistent...for UPS 1 it returns 100% (the capacity), for UPS 2 it always returns the load (18%) and for UPS 3 it always returns the battery temperature (28 - 29 degrees Celsius).2-4-2014 3-15-52 PM.png

0 Kudos

Are you getting these results in actual alerts or only when you test fire the alert from Advanced Alert Manager?

Testing alerts doesn't always test against valid UnDP data. Generally, the test will send the results of the first data point it finds for the object; the real test is of the alert action (email, snmp trap, etc). If you are seeing valid data on your linear gauges, then I would guess that you are getting solid information in your polling.

You can confirm with this SQL query:

SELECT

  n.caption AS 'DEVICE',

  cpa.assignmentname AS 'UnDP',

  cp.rate AS 'RATE',

  cp.rawstatus AS 'RAW',

  cp.status AS 'STATUS',

  max(cp.datetime) AS' LAST POLLED'

FROM

  CustomPollerAssignment cpa

JOIN

  Nodes n ON n.NodeID = cpa.NodeID

JOIN

  CustomPollerStatus cp ON cp.CustomPollerAssignmentID = cpa.CustomPollerAssignmentID

WHERE

  cpa.assignmentname like 'upsAdv%'

GROUP BY

  n.caption, cpa.assignmentname, cp.rate, cp.rawstatus, cp.status

-ZackM

http://www.loop1systems.com

Thanks for the clarification on the test firing. Initially, the UnDP was returning the incorrect value through normal polls, but of course I used the test fire to troubleshoot the issue, which I now know does not work in this situation. After confirming the pollers were looking at the correct nodes and values, it seems to have resolved itself. Everything is copacetic. Appreciate the assistance.

0 Kudos

Alright, Jim...you holding out on us? Where'd you get the cool the cool graphics? You know you wanna share your gift with the rest of the class. Please and thank you.

D

0 Kudos

Sorry deverts, meant to reply weeks ago. As blsanner mentioned, these gauges are built-in. Just add the desired gauge as he indicated and then edit the resource, and pick the appropriate UnDP that you created to associate to the gauge. If you need to reverse the gauge, for example, in the case of battery level you want 10% to be critical rather than 90%, click the "Show" button beside the Advanced section header (on the resource config) and then select "Reverse Thresholds".

pastedImage_0.png

View solution in original post

0 Kudos

jbartlett and blsanner,

I think that's where I went wrong, I didn't use the "linear" resource, so I have speed dials instead. Oh, well, maybe I'll change it.  Thanks for pointing out the error of my feeble ways.

D

0 Kudos

These gauges are built in.  You can add them by going to Customize Page --> Add Resources --> Universal Device Poller - Linear Gauge.  Then edit the resource and choose which gauge you want to use.  The three above are, respectively, Black LEDs, Black Round LEDs, and Black Dial.

0 Kudos