Performance threshold based on expression of SNMP OIDs

[Rob Eagle]

Rob Eagle [http://community.zenoss.org/people/reagle] created the discussion

"Re: Performance threshold based on expression of SNMP OIDs"

To view the discussion, visit: http://community.zenoss.org/message/74089#74089

--------------------------------------------------------------
Jane,
I know this is an old post, but did you ever find a solution for this?  Have a similar issue with trying to get the avg latency for cisco's ip sla avg latency.
--Rob
--------------------------------------------------------------

Reply to this message by replying to this email -or- go to the discussion on Zenoss Community
[http://community.zenoss.org/message/74089#74089]

Start a new discussion in zenoss-users at Zenoss Community
[http://community.zenoss.org/choose-container!input.jspa?contentType=1&containerType=14&container=2003]

[jcurry]

jcurry [http://community.zenoss.org/people/jcurry] created the discussion

"Re: Performance threshold based on expression of SNMP OIDs"

To view the discussion, visit: http://community.zenoss.org/message/74252#74252

--------------------------------------------------------------
I have just installed the community ZenPack, ZenPacks.zenoss.CalculatedPerformance.  This resolves the issue beautifully when you want to create a threshold that is an expression of several other datapoints.  Indeed, you can have an expression that is a mixture of other datapoints and of values on an object.

For information, go to http://wiki.zenoss.org/ZenPack:Calculated_Performance http://wiki.zenoss.org/ZenPack:Calculated_Performance .

I have added a comment to this page which may save you some of the hours I wasted.  The dialogue where you define your expression is a scrollable window.  I didn't notice.  I entered my expression into the Description box (which does nothing) and didn't scroll down to fill in the Expression box which DOES do something.

Other than that, it works a treat for me - both on device templates and component templates.  Here is an example from the readme.  Basically you create a new datasource (which creates a new datapoint) and you can then use that in thresholds and graphs as it does have its ownn rrd file.

For example, (hw.totalMemory – memAvailReal) / hw.totalMemory
The above expression uses the totalMemory attribute modeled from the device and uses the RRD data point memAvailReal to calculate a percentage used.

Cheers,Jane
--------------------------------------------------------------

Reply to this message by replying to this email -or- go to the discussion on Zenoss Community
[http://community.zenoss.org/message/74252#74252]

Start a new discussion in zenoss-users at Zenoss Community
[http://community.zenoss.org/choose-container!input.jspa?contentType=1&containerType=14&container=2003]

[Rob Eagle]

Rob Eagle [http://community.zenoss.org/people/reagle] created the discussion

"Re: Performance threshold based on expression of SNMP OIDs"

To view the discussion, visit: http://community.zenoss.org/message/74265#74265

--------------------------------------------------------------
Jane,
Sorry you did the leg work on testing this, but am very greatful for the it.  I believe this will close the gaps for many of us dealing with this issue and what a useful zenpacks.
Hope you are enjoying your afternoon and thanks again for the help,
Rob
--------------------------------------------------------------

Reply to this message by replying to this email -or- go to the discussion on Zenoss Community
[http://community.zenoss.org/message/74265#74265]

Start a new discussion in zenoss-users at Zenoss Community
[http://community.zenoss.org/choose-container!input.jspa?contentType=1&containerType=14&container=2003]

[tzel]

tzel [http://community.zenoss.org/people/tzel] created the discussion

"Re: Performance threshold based on expression of SNMP OIDs"

To view the discussion, visit: http://community.zenoss.org/message/74380#74380

--------------------------------------------------------------
that looks great, thanks for the anwers and Jane thanks for your effort. :)

Ill have a look at this today and return here if I managed to figure it out.
--------------------------------------------------------------

Reply to this message by replying to this email -or- go to the discussion on Zenoss Community
[http://community.zenoss.org/message/74380#74380]

Start a new discussion in zenoss-users at Zenoss Community
[http://community.zenoss.org/choose-container!input.jspa?contentType=1&containerType=14&container=2003]

[Doug Syer]

Doug Syer [http://community.zenoss.org/people/dsyer%40nwnit.com] created the discussion

"Re: Performance threshold based on expression of SNMP OIDs"

To view the discussion, visit: http://community.zenoss.org/message/74755#74755

--------------------------------------------------------------
I find i usually can get away without using the calculated performance zenpack.  Not that anything is wrong with it but for us its another daemon to maintain, etc..

say if you want to threshold on interface errors at 1%, you can do this on the threshold for the interface errors:

(here.getRRDValue('ifHCOutUcastPkts')*.01)

so if the latest count of errors/5 minutes is > number of packets * .01

you can do an rpn also for the graph (I believe)

the only downside is that you cant include the current error % in a transform unless your transform grabs the value of outcast packets out of the rrd and that isnt somethign I would do unless you had a small zenoss deployment since I belive its expensive.  but you can do somethign like evt.summary = evt.message = "Interface error % is >1%, see graph for exact details"
--------------------------------------------------------------

Reply to this message by replying to this email -or- go to the discussion on Zenoss Community
[http://community.zenoss.org/message/74755#74755]

Start a new discussion in zenoss-users at Zenoss Community
[http://community.zenoss.org/choose-container!input.jspa?contentType=1&containerType=14&container=2003]