Lag + disconnect 1/2 instance EU.

Please run a WinMTR trace that is capturing while you are playing and captures the disconnection/lag problem.

I don't see anything in your PingPlotter to show any signs of problems, so either the PingPlotter wasn't running long enough (or at the time of the problem) or the intervals are so high that the stuff happened between them.

"

Reverenz a écrit :

Hello sir, I will try WinMTR but I remember having issue with this software, so we'll see.
By the way, how is 45% Packet loss in the 16th hop not a problem ?

It's not real loss. That's how.

If the packet loss were REAL loss, then you'd be missing packets on every hop after that one. What you are seeing is the node fail to return a ping response. This is common due to data mitigation techniques. They occasionally drop ping packets on some nodes because pings are low priority traffic and often used in DDoS attacks. So if the node has other more important things to do, it drops the ping packet.

What you're seeing is 1 node not returning a ping.. but as you can see, your last node returned every ping it was given (which goes through that same node that drops ping)... but it wasn't dropped. Because the intended target wasn't that node.

Just because you see some packet loss on one node, doesn't mean that node is necessarily dropping all packets... in this case, it's not dropping any where it matters.

"

Reverenz a écrit :

@Drakier
Ok, thank you for the clear explanation mate. I'm waiting for the new league actually so I wont be playing that much, maybe not at all. If I catch another derp moment while ingame I will definitely turn WinMTR on and I will update this post.

Make sure WinMTR is capturing BEFORE you experience the issue so that the issue can be properly captured. If you only start the trace AFTER the problem happens, it might not capture the full problem.

We generally recommend letting WinMTR run for a long duration because then it gives us a better idea of the overall health. The more data it captures, the more data points we get, and the more a "pattern" might emerge. Statistical analysis. Having only 3 data points for example doesn't give us a lot of variation that we can gather a pattern from. But having 300 points means we can see what's more "normal" (and by knowing normal, we can see abnormal).

Too much data in this case is better than not enough.