Diagnosing severe lag spikes (WinMTR logs included)

Hey good peoples,

I've been experiencing what seem to be severe lag spikes (sometimes 5 - 15 seconds long) for over a year. The symptoms are that my frame rate stays high, with all characters on the screen chilling out (visibly breathing, but not walking or otherwise moving), but the game is completely unresponsive for a period followed by a fast-forward effect where everything that happened during the unresponsive period occurs in rapid succession (aka rubber banding). If I'm just surrounded by normal enemies on a map, for example, spamming my mob-clearing attack will save me. If I'm fighting a boss or on a delve, it's pretty much guaranteed death and loss of XP.

I've tinkered around with WinMTR after reading the Primer in this forum section, but I don't how to interpret the results well enough (yet) to take any actions. Here are some recent results.

1) https://pastebin.com/QiD5DfT2
This is from last night, testing an instance I lag-died in (IP address taken from Client.txt). lag-82.ear2.Chicago2.Level3.net appears to show severe packet loss, and the next hop at ae-2-3601.ear2.Washington1.Level3.net shows even worse packet loss and a single ping spike of 3514 ms during this ~5 minute test. When does the PoE network layer begin? I think this is probably an upstream/downstream provider based on "We have no direct path to escalate issues if they're being caused by upstream providers (e.g. Level3, AT&T, Verizon)." (from the Primer thread). My local friends that play PoE do not seem to have this lag issue. Is that something I should take up with my ISP? I don't foresee that going well, but I haven't tried.

2) https://pastebin.com/Wz0e8a2V
I let it run while I was sleeping, testing that same instance IP. Those same two layers are showing huge packet loss and ping spikes.

3) https://pastebin.com/R9J2LP4b
I ran this one while I was at work today, testing the default US login gateway (Texas). I typically use this one or Washington DC, but I have the same issues with both. There's another Level3 hop with 100% packet loss and a max ping of 3601 (avg 762), followed by another hop with high packet loss (54%) and a max ping of 4989.

4) https://pastebin.com/iqyvgLtN
EDIT: I added this one the morning after the initial post. This is testing the Washington DC (US) login gateway from ~2 am to ~9:30 am CST, and again, there's a pair of Level3 hops with high packet loss and ping.


I also noticed what I'm going to pre-emptively call some "funny business" in the log when I was hunting for that instance IP. Here's the pastebin for that:

5) https://pastebin.com/c4BmiVh2
This message seems to be extremely common when I have these severe lagging issues, though it could be coincidental and unrelated for all I know: "Matching object found for InstanceClientActionUpdate, but no matching action was found on object to update."

So. Can one of you kind folks help me figure out how to resolve this? Did I skip over some obvious things you also need to see? My tired brain is losing focus, but I wanted to post this tonight anyway. Thanks in advance!

Edit: adding in that I'm on a wireless connection and have not tried any other computers. (but I think the WinMTR results above indicate my connection to my router isn't at fault here? emphasis on _think.)