I'm using a Flint 2 (GL-MT6000) running v6.2, with telegraf installed to monitor network traffic which is exported to influxdb and grafana running on a dedicated VM.
Monitoring eth1 (the interface with the public IP address from the modem) is only showing 1/5 of the incoming network traffic, but outgoing traffic on the same interface is correct.
Network acceleration is switched off.
I tested this by downloading 5 Linux Mint ISOs totalling 14.3G concurrently on a VM, and looking at the differences between Flint 2 incoming traffic, and VM's incoming traffic.
On the Flint 2, bytes_recv went from 92.63G to 95.58G.
The same thing happened before I upgraded from v5.8 (no settings kept during upgrade).
Has anyone else noticed similar behaviour?
I haven't set anything up in LuCI, as I understand that will rename my network interfaces and I'm not sure I want to go down that path as I would have to do it after every upgrade (correct me if I'm wrong here).
Network Acceleration is disabled, so that shouldn't be an issue.
But it's like the interface isn't recording all the traffic that's going through it, which is what I would expect to see if some sort of acceleration were being used.
And this is only for incoming traffic. Outgoing is showing the right numbers, using the same query.