I have a quite complex infrastructure in my house, and I monitor everything. Among others, I have SmokePing probes towards multiple destinations in my LAN, and also to my internet provider’s infrastructure, and even external hosts.
When I added a graph for the management IP of my Netgear GS108Ev3 8 Port Gigabit Ethernet Smart Managed Plus Switch, which is my “core” switch, I noticed that there is a huge packet loss, in the range of 4% average, about 40% max.
The strange thing is that other destinations in my LAN which can be reached via this switch, have absolutely zero packet loss. That means that the cabling itself is definitely ok, and also that the packet switching functionality of the switch is ok.
So I opened a ticket towards NetGear, and the response I got was quite surprising (but still somehow makes sense):
This is a known issue, and it is by design. The put the highest priority on the packets flowing thru the switch, while the management traffic has a lower priority, so that ICMP packets to the switch’s management IP may get lost, and in extreme cases the web UI may not be reachable at all.
It’s not a big issue for me, because in the end I wanted to create graphs that exhibit whether I have any internal packet loss. Whether the switch’s management interface has packet loss or not is not really important to me. As long as all my internal devices that I actively use, like my NAS devices or streaming boxes, are properly connected with no packet loss, all is well. 🙂