Just doing some digging on a couple of UBNT nanostations I’m running, it appears that UBNT are utilizing some SNMP OIDs from the Mikrotik MIB (enterprises.14988) in addition to their own (FROGFOOT, enterprises.10002) MIB. Most importantly, it’s the radio-side statistics OIDs… here’s what I’m getting from a client-side 3.65GHz radio:

FROGFOOT-RESOURCES-MIB::memTotal.0 = Gauge32: 29524

FROGFOOT-RESOURCES-MIB::memFree.0 = Gauge32: 11024

FROGFOOT-RESOURCES-MIB::memBuffer.0 = Gauge32: 2080

FROGFOOT-RESOURCES-MIB::memCache.0 = Gauge32: 0

FROGFOOT-RESOURCES-MIB::swapTotal.0 = Gauge32: 0

FROGFOOT-RESOURCES-MIB::swapFree.0 = Gauge32: 0

FROGFOOT-RESOURCES-MIB::loadNumber.0 = INTEGER: 3

FROGFOOT-RESOURCES-MIB::loadIndex.1 = INTEGER: 1

FROGFOOT-RESOURCES-MIB::loadIndex.2 = INTEGER: 2

FROGFOOT-RESOURCES-MIB::loadIndex.3 = INTEGER: 3

FROGFOOT-RESOURCES-MIB::loadDescr.1 = STRING: 1 Minute Average

FROGFOOT-RESOURCES-MIB::loadDescr.2 = STRING: 5 Minute Average

FROGFOOT-RESOURCES-MIB::loadDescr.3 = STRING: 15 Minute Average

FROGFOOT-RESOURCES-MIB::loadValue.1 = Gauge32: 0

FROGFOOT-RESOURCES-MIB::loadValue.2 = Gauge32: 0

FROGFOOT-RESOURCES-MIB::loadValue.3 = Gauge32: 0

MIKROTIK-MIB::mtxrWlStatTxRate.5 = Gauge32: 146250000

MIKROTIK-MIB::mtxrWlStatRxRate.5 = Gauge32: 146250000

MIKROTIK-MIB::mtxrWlStatStrength.5 = INTEGER: -60

MIKROTIK-MIB::mtxrWlStatSsid.5 = STRING: CHANGEDTOPROTECTTHEGUILTY

MIKROTIK-MIB::mtxrWlStatBssid.5 = STRING: CHANGEDTOPROTECTTHEGUILTY

MIKROTIK-MIB::mtxrWlStatFreq.5 = INTEGER: 3667

MIKROTIK-MIB::mtxrWlRtabStrength.'.'"FV.'.5 = INTEGER: -60

MIKROTIK-MIB::mtxrWlRtabTxBytes.'.'"FV.'.5 = Counter32: 2031561417

MIKROTIK-MIB::mtxrWlRtabRxBytes.'.'"FV.'.5 = Counter32: 704139062

MIKROTIK-MIB::mtxrWlRtabTxPackets.'.'"FV.'.5 = Counter32: 111296

MIKROTIK-MIB::mtxrWlRtabRxPackets.'.'"FV.'.5 = Counter32: 151967791

MIKROTIK-MIB::mtxrWlRtabTxRate.'.'"FV.'.5 = Gauge32: 146250000

MIKROTIK-MIB::mtxrWlRtabRxRate.'.'"FV.'.5 = Gauge32: 146250000

Looking at the AP side walk, it shows RF data for both of the connected stations on this one AP, but the differentiator isn’t using a normal SNMP OID number, rather it’s a field, in my case “FV.” vs “FW.”…

MIKROTIK-MIB::mtxrWlRtabStrength.'.'"FV.'.5 = INTEGER: -64

MIKROTIK-MIB::mtxrWlRtabStrength.'.'"FW.'.5 = INTEGER: -70

So that may make things a bit more interesting trying to figure out how to poll/graph that information.

I haven’t even tried beating on any of our AF24s yet, that’ll be next.

…Ron

From: observium [mailto:observium-bounces@observium.org] On Behalf Of Keefe John
Sent: Tuesday, October 28, 2014 9:49 PM
To: Observium Network Observation System
Subject: Re: [Observium] Wireless signal graphs

Adam,

I would love to see support for Airfiber and Ubiquiti Airmax radios too.

Thanks!

Keefe

On 10/28/2014 8:34 PM, Justin Miller wrote:

Adam, 

Unfortunately I think the wireless ISPs have figured out how good this product is and you are only going to get more requests for this kind of thing.

When the time comes, I can make available wireless hardware for Ubiquiti AirFiber 24 ghz and 5 ghz links as well as Mikrotik point to point and point to multipoint links for use by the development team. (public ip snmp access) Also willing to lend wireless experience/explanations.

Ubuiquiti’s snmp implementation on the AirFiber’s are just wacky. Only traffic that the CPU processes shows under the usual interface spot i.e. management traffic. The traffic you actually want to see is in a whole other section. It’s mostly because of the hardware structure of the radio with is dsp’s and such.

Justin

On Oct 28, 2014, at 6:17 PM, Adam Armstrong <adama@memetic.org> wrote:

The SNMP implementations and MIB structure on these wireless devices make me want to stab out my eyes.

We have some vague plans to try to work out some infrasdtructure to fit wireless kit into in the future, but none of us have any wireless experience, and each of the vendors seem to have approached the problem in (at least one) entirely different (and usually equally broken) manner, making it even harder to come up with a common schema.

At the moment there isn't a place to put bounties, mostly because it encourages people to expect major (and often ill-advised) changes for small amounts of money. ;)

adam.

------ Original Message ------

From: "Travis Kreikemeier" <travis@travko.com>

To: "Observium Network Observation System" <observium@observium.org>

Sent: 10/28/2014 10:54:05 AM

Subject: [Observium] Wireless signal graphs

Any plans to add wireless signal metric graphs to Observium?  Such as signal strength and SNR.   Example device being Cambium Canopy CPE’s.

Is there a place to put feature bounties and payments for completion bounties?

Thanks

TRAVIS KREIKEMEIER

_______________________________________________
observium mailing list
observium@observium.org
http://postman.memetic.org/cgi-bin/mailman/listinfo/observium

_______________________________________________
observium mailing list
observium@observium.org
http://postman.memetic.org/cgi-bin/mailman/listinfo/observium