It seems #2066 is not fixed. We’re experiencing a similar problem : the average traffic is completely different than the average measured.
In the screenshot below it is obvious that the sum of the average in+out is 1.80 + 2.11 = 3.91kbps but in the usage it is listed as 31.20kbps, almost 10 times as high.
This happens on all our devices.
Does anyone have any idea what might be wrong ?
But your total bandwidth indicates that number looks correct. Take total bandwidth divided by time passed.
I think you are missreading that value. Usage is in 95th percentile and not down+up avg.
Billing was not set up to ‘CDR 95th’ but to ‘Quota’, s 95th percentile shouldn’t have anything to do with this ?
In any case it doesn’t make sense. This is the billing graph for 1 month (which is less than a month, since it was set up recently) :
It shows 3.86GB usage.
But if I look at the actual graph of that same interface :
It shows 2.71GB usage, which is mostly explained by that big peak around the 29th. That peak is before the billing usage period started.
Well. Usage are both: transferred data (95th percentile)
In my case I use it to “control” what one of our transit bills to us. They use cacti to monitor this.
I was going to say that theirs match with ours but it only does in percentile. In transferred data LibreNMS shows like 70TB more than our transit provider graph…
773.510TB LibreNMS VS. 702.8TB Cacti <-- Transferred Data
3.780 LibreNMS VS 3.754 Cacti <-- 95th Percentile
So yes, something is wrong here.
PS: Also noticed that Percentile is not doing inbound+outbound, only inbound.
You guys need real numbers to show that there is an issue. Data from the graphs is not enough.
What do you need in terms of numbers ?
I don’t need anything. You need to show what the numbers should be and why. Perhaps manual measurements, I dunno.
I showed the real numbers. The ones that my transits said we did 
Sorry, I don’t know if I was miss-understood. I don’t know anything about billing, I can’t help you. I was just trying to guide you in the right direction. Probably next you need to inspect the billing and see what it is actually measuring and how it calculates it.
Its ok. I’ll try to dig into it. But its not only billing. The data transferred in port graph is also wrong, so it must be something shared.
This is what the port reports:
That smells like they aren’t counting the same thing (or in the same units)
And you said it!
Cacti is showing what now is called Tebibytes(which used to be Terabytes in base 2), LibreNMS is showing the new base 10 Terabytes.
WTF should we do now?
You can convert the numbers from one to the other. Do they match?
I directly did from the bytes in the database.
From bytes to Tebibytes, its 703TiB, almost exact as our transit graph, to Terabyte its what Libre shows.
@wimg
Try this in your config.php
$config['billing']['base'] = 1024;
