Path of Exile 2 — Network Spikes After Patch (0.3)

"

DcKorin#0208 schrieb:

Changing networking type seems to be it for me + picking the lowest latency loginserver.

- poe2_production_config.ini file, change networking_type=auto to networking_type=predictive


Before that, it was possible but stupid workaround: die quick ingame until you get a good and clean zone with your respawn. Sometimes 1-3 deaths were enough, sometimes 7-10 to get it. You can see it immediately on the networking graph if it's ok...

"

BallinGorilla#5889 schrieb:

Region/Gateway: Canada / Washington DC (tested multiple)
System Specs:

AMD Ryzen 5 7600X

32 GB DDR5 (6000 MHz)

NVIDIA RTX 3060 Ti FE

Windows 11 (64-bit)

Wired Ethernet (Bell ISP, Canada)

Issue Description

Since the most recent 0.3 patch, I’ve been experiencing severe and consistent network latency spikes in Path of Exile 2.

Before the patch → game was smooth on Canada East (Toronto).

After the patch → frequent latency spikes shown on the in-game F1 graph, causing rubberbanding/freezes.

FPS is stable (120–140), CPU/GPU temps are normal, no local performance issues.

Troubleshooting Performed

✅ Tested on both Canada and Washington DC gateways → spikes persist on both.

✅ Switched Networking Mode (Lockstep ↔ Predictive).

✅ Lowered graphics to minimum → FPS increases, but latency spikes remain.

✅ Disabled V-Sync, used Fullscreen, tested Vulkan vs DX12.

✅ Disabled Antivirus/Web protection during test.

✅ Tried via VPN (New York/DC exit) → spikes still present.

✅ Local pings (router, 1.1.1.1, 9.9.9.9) are stable.

✅ System resources checked → no CPU/GPU bottleneck, temps ~70 °C.

Conclusion

This does not appear to be a hardware or local network issue. Since VPN routing does not bypass the spikes, it suggests the problem is upstream from multiple ISPs or within GGG’s server infrastructure after the recent patch.

Request

Can GGG confirm if there are known issues with the Canada East and Washington DC gateways since patch 0.3?

Any workarounds or recommended testing beyond what’s been done?

Screenshots of F1 latency graph (spikes) and Resource Monitor logs available on request.

"

szabarna1000#4165 schrieb:

The only thing that actually worked for me was

- Disabling global, trade chat
- Setting sound channel count to low