Too much skill lag on reset night, instant casts don't even go off.

[Kelly.5293]

Too much skill lag, instant casts don't even go off on SF. My guild of 40 just dropped after 30 minutes on Friday reset. It's unplayable and a waste of time.

No idea what's going on.

Sadface

 

[mysticozzy.3589]

*in before someone says the system cant handle it* 

which is true the servers are kitten and grossly outdated

[kamikharzeeh.8016]

its the same since many years, even gotten worse when they swapped to amazon servers tbh

funnily its random, between players too. i'm EU timezone and lag less in NA resets often than in EU ones.

and the key is still: spam your stuff faster than enemy and don't try to get the full lagged things off... (as they wont get off anyways as logn as u are in highlag territory)

idk what u refer to with "SF" tho... skillag isn't only one one specific server or timezone. if it happens, most people have it. better PCs fare a bit better, better web ofc too.. but everyone can and will have it in blobsize threesome fights, kinda inevitable as its anet servers lagging there

there are no real ways around it so ye

[TheGrimm.5624]

Mind you not saying no and not an Amazon fan in anyway, but did you do any traceroutes? Didn't check on global DDOS BS Friday night but sadly that is typically seems to be a happy time for those ..... (yeah don't need infracted) peeps. Had guildmates on different parts of the NA as well running into differing issues while others did not. Which leans toward less server and more points between. Its annoying all the same and that does suck.

[Dawdler.8521]

I didnt have any lag during reset 👍

Still logged within 30m because my guild of 2 people represented 66% of the population on the border.

[TheGrimm.5624]

17 minutes ago, Dawdler.8521 said:

I didnt have any lag during reset 👍

Still logged within 30m because my guild of 2 people represented 66% of the population on the border.

 

Ouch on the matchup.

That's what made me question interconnects. I have a terribad internet where I am at and a friend having an issue has a great one. Which meant though not a fan of the company-provider, it might be more of something between the points of connection. 

[DeceiverX.8361]

It's caused by the calculations being done for AoE's and boon/condition stacks/durations based on what has been announced by developers in the past.

More targets being hit by more stuff creates more lag.

Which is why the increased emphasis on companion AI is bad news for WvW.

[TheGrimm.5624]

8 hours ago, DeceiverX.8361 said:

It's caused by the calculations being done for AoE's and boon/condition stacks/durations based on what has been announced by developers in the past.

More targets being hit by more stuff creates more lag.

Which is why the increased emphasis on companion AI is bad news for WvW.

 

I could see that, but that wouldn't explain why player A and player B didn't see the same issue in the same fight while being geographically in two different locations. If it was all server side both players would encounter the same impact. Which has been seen and would match what you refer to. Latency exists, no doubt, I dislike Amazon, will have to trust me there. But there are also issues between connections. Hop 1 and hop 25 may vary quite a bit. If its not on the ends, but in the middle its not something the player, Amazon or ANet can control. If we find its server side, then indeed ask ANet to tell Amazon to get its act together. hence any data on what was seen?

[Aeolus.3615]

On 11/12/2022 at 3:55 AM, mysticozzy.3589 said:

*in before someone says the system cant handle it* 

which is true the servers are kitten and grossly outdated

It is not only the server but theres several "soft" bottlenecks in alot of places,  imagine has if the server is bottlenecked due client relationship and client is bottlenecked due the ammount that CPU can handle, therefore theres a queue of data and some data will be lots as in skills that wont make effect or game validation like walls for example that wont reach the client, ive ended 2 times  with invisible stucture walls, the enemy players in their client would see me get trough walls of their keeps (a simple game restart will fix this).

When we have skill lag on small combat scale thats Anet server  load balancer borking up for sure.

The only decent playtime i had was with a cpu OC'ed @ 6GHZ on a dedicated fiber 1GB down 500MB up, was running everything @max with stable frame rate in wvw 3 way queue.

I am actually expecting that new the cpu's  with high clock will solve a bit the bottleneck feeling the game has, mmo's are heavy single core junkies so far, the better and faster the single thread performance is  the smooth the game runs.

Edited November 14, 2022 by Aeolus.3615

[DeceiverX.8361]

20 hours ago, TheGrimm.5624 said:

 

I could see that, but that wouldn't explain why player A and player B didn't see the same issue in the same fight while being geographically in two different locations. If it was all server side both players would encounter the same impact. Which has been seen and would match what you refer to. Latency exists, no doubt, I dislike Amazon, will have to trust me there. But there are also issues between connections. Hop 1 and hop 25 may vary quite a bit. If its not on the ends, but in the middle its not something the player, Amazon or ANet can control. If we find its server side, then indeed ask ANet to tell Amazon to get its act together. hence any data on what was seen?

 

This depends so heavily on how the netcode is written that I don't think any observable data by any small grouping of players over a limited amount of time will actually provide meaningful sampling.  If based on hops alone, the lag would be bad and stay bad no matter the size of the engagement for the player in all environments, not just WvW.

If a server's event handler is struggling to keep up, there will always be some requests that get done first and are thus responsive.  If this is a distributed system, it could be that some threads are simply locking others out while consensus is trying to be reached, but due to timeouts, the faster-solution threads are already re-calculating new data and ignoring consensus from others as to avoid causing the whole system to get bogged down trying to force consensus and stalling them all.  This is a somewhat crude but effective way or keeping the most amount of people lag-free for the most amount of time.

We just don't know.  What we do know is that there is an observable phenomenon where players not in the same fights experience different levels of this sort of buffering, likely as a result of proximity, suggesting the processing and respective multi-core programming is dynamically based on location data of concentrations of players in some capacity, which may also explain phenomenons involving culling in their legacy code.

I'm going to inherently trust the statements from their developers about it being based on too many calculations based on targets and conditions and boons in AoE settings more than any amount of player speculation with the limited data we have; frankly, while I do write a ton of embedded code for a living, I'm absolutely certain the network engineers and multiprocessor specialists at ANet and those at AWS are significantly better at writing it for such large-scale distributed processing than I.

Edited November 15, 2022 by DeceiverX.8361

[Daniela.6285]

A lot of people complaining about lag. Impossible to play in ebg NA 

[Karagee.6830]

Change server. Go to servers with fewer people. Not outnumbering your opponents and being at map cap every time you play, may be a new experience for ya.

Edited November 17, 2022 by Karagee.6830