Odd spikes in SNMP reports

Hello friends!

For the past few months I've been noticing some odd behaviour in the graph of my SNMP-enabled router. It will occasionally spike to just above 800MB at random intervals for a second or two and then return back to the actual normal numbers. This is annoying because my connection to the internet can't go above 13MB and it renders the whole graph effectively useless until the spike moves along over time where the scale returns to normal. Example attached.

Because I have quite a long graph this takes some time.

I've only really noticed this in the most recent version (or two) but have just been putting up with it. Has anyone else experienced this at all? If not, what additional information would help?

Thank you in advance!
Zar

From what I can see on your screen shot, the time between the spikes is about 400 seconds, which at 10 MB/s would give us around 4 GB.

Since the standard 32-bit SNMP counters roll over every 4 GB (2^32), the spikes seem to be happening when the rollover takes place. It could be a bug in the router's firmware, or a bug in NetWorx, and we will need to investigate this further.

If you are using a version lower than 7.1.3, please update to the latest version and check if the issue persists.

Otherwise, could you please tell me the model of your router to get started?

Hi Andrew,

Thanks for replying!

Okay, here's the details of my router:
TPLink Archer VR2100v v1
Firmware Version:1.13.0 0.9.1 v009e.0 Build 220623 Rel.61658n

Also, I'm using 7.1.3. And I've made sure that the firmware on my router is the latest version.

Let me know if you need further info.

Thanks!

Thank you, we have prepared a debug build that will log some numbers from the router, so we can see what is going on.

Please update your installation to the latest build from here. Upon launch it will create two text files on your desktop, named comm.log and diff.log.

Then simply wait for the issue to occur, i.e. an erroneous spike on the graph. It will occur sooner if you download a large file or similar, transferring over 4 GB. Finally once the spike has manifested itself, please compress the logs to a ZIP file and send it to us.

We will then be able to see what data the app received from the router and how it handled it, hopefully shedding light on this issue.

Hi Andrew,

Thank you again! I've just upload the logs via message as you requested.

Thank you, we have checked the logs and it appears to be a bug in your router's firmware. Here is the relevant section from the log:

12/09/2023 5:38:24 PM in num = 4276320451 out num = 2425539757 in fmt = 4.0 GB out fmt = 2.3 GB
12/09/2023 5:38:25 PM in num = 11776129 out num = 2425717128 in fmt = 11 MB out fmt = 2.3 GB
12/09/2023 5:38:26 PM in num = 8867592 out num = 2425908088 in fmt = 8.5 MB out fmt = 2.3 GB
12/09/2023 5:38:27 PM in num = 22569110 out num = 2426109827 in fmt = 22 MB out fmt = 2.3 GB
12/09/2023 5:38:28 PM in num = 36254912 out num = 2426306365 in fmt = 35 MB out fmt = 2.3 GB

Normally the "in" and "out" counters steadily increase except when they roll over. However, in your logs, following the 4GB rollover, the router first reports that it received 11 MB, then 8.5 MB, then 22 MB, then 35 MB, and so on. The issue is that the value reported at 5:38:26 PM is actually lower than at 5:38:25 PM, which resulted in a miscalculation and the spike that you saw.

In the new version, we have added a workaround that will simply discard erroneous data like that.