Desync whenever I die, or use a movement skill rubberbanding after 10-20sec resyncs me

[Pinkbunnyman]

Connected instances:

ec2-52-4-240-116.compute-1.amazonaws.com:443 - Assume this is the aws load balancer for multi instances
ec2-35-166-55-23.us-west-2.compute.amazonaws.com:443 - The login server.
ec2-35-157-63-228.eu-central-1.compute.amazonaws.com:443 - Assume this is the game server by traffic volume
85.195.95.29:443 - seems to be where the ping comes from; which is NOT the same server as the game server.

Tests performed:
Constant ping to 8.8.8.8, max jitter of 100ms over 500 pings (while dying to create the desyncs)

ec2-52-4-240-116.compute-1.amazonaws.com:443 - Average latency 148, with jitter of 400 min of 90 and max of 490ms

ec2-35-166-55-23.us-west-2.compute.amazonaws.com - Average latency of 316 jitter of 405, min latency 153 max 558

ec2-35-157-63-228.eu-central-1.compute.amazonaws.com - Average latency of 314 jitter of 408, min latency 150 max 558

85.195.95.29:443 - No response to pings, however RTT of 30ms (same as in-game’s “ping”)

In-game suggests my latency is 30ms which doesn’t seem to be the game server (which explains the post of lag at low latency)
EU-Central seems to be the AWS cost for the “EU-West” server cluster, I’d suggest moving this to EU-West-1 or renaming the server as this could cause internal confusion in the future.

My latency to EU-Central-1 (Frankfurt) is just over 30 miliseconds, so I believe that the network throughput of the EC2 clusters isn’t sufficient, I’d suggest contacting your account manager and seeing what can be done in terms of upping your vcpu limit so you can reduce the maximum jobs on each EC2 instance to maintain network integrity.

Edit: I’d also suggest adding instance connections to your logging at the moment, so you can see which instances are causing issues, it will also let you know if there’s an issue with one specific cluster

[Pinkbunnyman]

Also in my player.log I have

2023-03-14T00:02:44.1232360+00:00	Cosmos latency warning, https://lastepoch-data-prod.documents.azure.com/dbs/lastepoch/colls/characters/docs/3E76F4502C95E5C3 GET responded with code 200 and took 662.1881 ms (activity ID 39598ef9-883e-4987-ad84-3ab78cfaa3d4)
LE.Telemetry.PrependTimestampLogHandler:LogFormat(LogType, Object, String, Object[])
UnityEngine.Logger:Log(LogType, Object)
UnityEngine.Debug:LogWarning(Object)
AzureCosmosClient.<CallApiAsync>d__16:MoveNext()
BeforeDeathAction:Invoke()
Cysharp.Threading.Tasks.Internal.PooledDelegate`1:Run(T)
System.Action`1:Invoke(T)
UnityEngine.AsyncOperation:InvokeCompletionEvent()

This is unrelated to my rubberbanding issues, this appears to just be another symptom, however, if there is too much latency it could result in failed or slow logins