How to monitor live throughput?

Hello everyone! My name is Zack, and I'm a big part of my county's largest WISP. I'm young, and fairly new to the field and lack official traning. I've been tasked with taking the company and charging forward with it. So I'm here looking for a bit of guidance and any sort of help that I can get.

We recently upgraded two of our networks backhauls from Ubiquiti Rockets to ePMP 1000's. Throughput definitely seems to have gone up at both locations and overall I'm really happy with the upgrade. However, in the weeks since the upgrade I've found myself wishing there was a traffic graph in these radios to monitor real time throughput, one of the few things I really liked about the Ubiquiti radios was the realtime traffic graph. 


I was wondering how/if anybody has an alternative solution for monitoring live throughput on these radios. 

Apologies ahead of time if this is an ignorant question, I've got a lot to learn. 


-Zack

Hey Zack, welcome to the forums!

In earlier ePMP firmware releases there was a traffic graph, but it caused some UI issues and they removed it recently. There are plans to bring a traffic graph back. In the mean time your best bet is to use a managed switch to watch traffic on the port that the ePMP's are connected to, OR use SNMP to grab the stats off of the ePMP radio and use a monitoring/graphing tool to display the data.

Eric is spot on.

There used to be a live graph and it went away in 2.4.x firmware. They are talking about brining it back, but there was a memory leak that caused them to remove it.

If you're using a monitoring program that can trend SMNP that would be a good work around for now. We have several thousand devices (not customers) in TheDude.  It has nice graphing capabilities.

Can you describe how you're monitoring throughput on ePMP with the Dude?

Hi Zack. If you do figure out how to effectively monitor live throughput in real time, please post it back here. We have a bunch of ePMP gear, but we haven’t been able to deploy it live, partly due to the inability to see the instant client’s throughput live.

So, if you do figure out a way to see all connected clients and their throughout in real time, please let me know. :slight_smile:

I know that at some point in time prior to my joining the company "The Dude" was used for monitoring, but has since gone to the wayside. 

I'll definitely need to do some research on it and see what we can get set up with it.


Wonder how long it'll be before Cambium gets a solution... I'm not a big fan of the GUI in the ePMP radios. It seems to lag a lot, and occasionally lock up all together.

The dude will do the job. SNMP can take sometime to setup but is well worth the effort. Not only can you monitor it real-time, you can retain that information and look at what your network was doing days, weeks or months ago. You can setup alarms to alert you when through put hits X to let you know when a link is getting full and needs capacity added, it can track device heath parametiers such as temprature, cpu and memory loading on compatiable devices We’ve had whatsupgold since day one, which is essentially the same thing but a paid software. We store 4 years of history in what’s up gold to aid us in future planning. Unfortunately not all devices can go so in-depth as the health monitoring (ubnt supports very little SNMP) but Cambium hardware and most other vendors do support plenty of outputs and traps for you to automate most of your management loading.

We're currently routing with a PFSense router that keeps a lot of great detailed statistics but only a collective snapshot of the entire branch of the network at a time. It seems difficult to get more detailed specific information with it. The Dude definitely sounds like what we need. Time to start studying! 

1 Like

Hi. I’ve looked at The Dude and I don’t see anywhere that it can show a Cambium AP and a real-time display of the throughput of the clients. But, I don’t know it that we’ll, so hopefully I’m just missing something? I’d be thrilled to learn that TheDude can do that. :slight_smile:

Hi.  Is there a way to access the information that is on the AP -> MONITOR -> PERFORMANCE list from the CLI, or via SNMP or somehow to get at this information off this webpage in some way?

Basically, we need to be able to monitor the clients connected to an Access Point, and to be able to see if they are transferring data, and to see how quickly they are transferring data.  This list gives me almost all of that info - since it does have the traffic counters moving upwards. However, it's pretty impossible to do the math as the data is changing and to somehow calculate a throughput rate from that.

So, if for example I have "CONFIGURATION -> SYSTEM -> WEBPAGE AUTO UPDATE" set to 5 seconds, then all that really needs to be done to calculate a real time throughput rate is to take ''total kbit current value, and subtract total kbit previous value, and divide by 5 seconds'' and I'd have a throughput rate.

So - ideally it would be nice if this screen was modified to do that; to have another couple columns that does the math ''current total kbit - previous total kbit / auto update interval'' and display the throughput rates.

If that can't be accomplished for some reason, then is it possible for me to get this same list of information via the CLI or is it possible for me to get the Webpage Source Code so that I can try to modify it so that it can show us a live throughput monitor?


Thank you. :)

Hi,

I suggest you to use CNS server software which is a monitoring & managing software tool for ePMP devices.

Below is the link to download CNS server software.

https://support.cambiumnetworks.com/files/cns%20server/

Using CNS we can manager 1000 devices. Also this will be helpful for real time monitoring. It has dashboards under monitoring section. For more information on CNS, please go-through the user guide and release notes available in the above link.

I hope this will meet your requirement. Please let me know, if this is helpful.

Regards,

Vijay Gnanamurtthi

Why reccommend CNS Server over cnMaestro?

 
Neither will not show realtime values ​.
Neither will not show realtime values ​.

OK, thanks Hans.  That's what we're discovering too with every sugestted solution. They are usually 5 minutes averages, so not real time.

The frustration is that the Cambium ePMP Access Points do already know this info.  In the AP under MONITOR -> PERFORMANCE, it is already displaying the number of bits flowing in & out every 2-20 seconds (depending on the value in CONFIGURATION -> SYSTEM -> WEBPAGE AUTOUPDATE).  We set ours to 5 seconds, and that's great.

So, the ePMP Access Point already DOES display nearly real-time throughput which is great, but it's displaying total 'Total Kbits' and it's just not subtracting that from the previous total so that it could display 'Kbits since the last autoupdate', which would be MUCH more useful information.  Or better yet, dividing that delta by the autoupdate interval in order to show Total Kbits/Second over the last 5 seconds.

So, hopefully Cambium can add in a 'KBit Delta' column, or a 'Ave Kbit/Second' column.

+1 on adding 'recent' throughput to session/link stats in numeric form.

Last I remember reading here, Cambium still intends to restore the live throughput graph feature that was present in earlier firmware versions.  The information it's based on, as you pointed out, is already present in the status update the javascript UI pulls via AJAX every x seconds.

j

1 Like

To jump in here...  We were after the very same thing.  We use cnMaestro from time to time and it is decent for graphing but we needed monitoring as well (with notifications if APs or SMs went down).

After looking for inexpensive (free) options, we settled on PRTG. It does need a Windows machine to run on but was very easy to set up and can monitor everything.  We do SNMP of our switch ports, router and APs.  We just ping SMs but you could also get SNMP data from each SM.  You can then set up alerts if traffic on any device goes above certain thresholds, etc.  Very customizable.

We tried Nagios on a Raspberry Pi in our NOC but it was going to take a degree in annoyance to get it running.  We also tried CNS, it was OK but looked like it was written in the 80's.  Spiceworks.com is also good - very pretty and easy to set up but not quite as customizable as PRTG.

If I rember well, if using StarOS as a router, you can see traffic by ip address via the Beacon Realtime Traffic Monitor. It can also be done in Mikrotik and sure their a way in Edgerouter via CLI and/or debian paquages.

We have CACTI servers, MRTS server, PRTG servers as well as using the cnMaestro product - and they all are useful and all do their own things quite well.  None of them show the real time throughput - they all show historical throughput. You can see updates as much as 5 minutes ago - but nothing that shows what's going through now, or in the last 5 or 10 seconds.

That info is 'almost kinda' already available right in the GUI's, since it's showing kbit total updates every refresh - so if you set your webpage autoupdate interval to 2 seconds or to 5 seconds, then it is already updating the in/out data totals every 2 or 5 seconds.  It's just not subtracting the previous value from the current value so that we can see how much data (or what speed) that happened in the last 2 or 5 seconds.  :(

Hello

We are using Observium at our Data Center for over 3 years now to monitor every device health and traffic.

http://observium.org

We are also monitoring our radios ePMP, and Mikrotik routers. 

Somehow ubnt is on snpm v1 yet, so there is not much information we can pull out, but Cambium works great.

Here is an example graph

Cambium ePMP 1000

Cambium ePMP 1000

Mikrotik Router

 Live data is possible with PRTG as well.  

 This link is a view to our live data on one of our backhauls.

1 Like