Device Interface Statistics

[mattg]

Can device interface statistics be queried through the API? I am looking to obtain bandwidth stats on uplink interfaces. (in kbps) SNMP doesn't offer specific bandwidth usage stats in kbps, and doesn't offer a time value (like uptime) to accurately calculate bandwidth.

Thanks

[MacuserJim]

The only statistics I know about being able to get via API is what interface are being used and the amount of sent and received data.

[Nolan H. · Accepted solution]

Would netflow give you what you want?

[mattg]

>Would netflow give you what you want?

There's a lot of overhead and excess data for which I don't really care about in Netflow. As well, Netflow is enabled for the whole site. Not just an interface; or a specific device. It's enabled "network wide". I don't know what that means. I have a firewall, switches and several wireless access points. Am I receiving an aggregate value of all devices?
That being said, I am trying to use Netflow, now. My system has the ability to collect Netflow data. Though that has its limitations too. At least I have a better insight into traffic utilization.

Thank you

[Nolan H.]

Netflow traffic will only be coming from the MX per your network you have it setup for. But I understand what your saying 😃

https://documentation.meraki.com/MX/Monitoring_and_Reporting/NetFlow_Overview

[mattg]

> Netflow traffic will only be coming from the MX per your network you have it setup for. But I understand what your saying 😃

Thank you, Nolan. I also found something which alluded to NETFLOW being sourced from the MX, only. Which is fine...for me...because that's the uplink I wanted to monitor, anyway.
I'll also say...despite my apprehension to using NETFLOW...it's already turned out to be a much more useful source of information. As suspected, I discovered some bandwidth anomalies; and with NETFLOW, I was able to determine the source very quickly.

[Philip D'Ath]

If you query the interface data volume every 60s then you can calculate the average bit rate. You don't need to worry about the timing on the device - you just need to time when you make your queries.

[mattg]

> If you query the interface data volume every 60s then you can calculate the average bit rate. You don't need to worry about the timing on the device - you just need to time when you make your queries.

For a brief, manual-effort analysis, that would be adequate. As part of an on-going, systems monitoring process, that may not be very accurate method. There's overhead and latency on the system performing the queries. (and I'm not sure that I could even have the platform do the math.)
Thank you for the reply, though.