Persistent server-side latency accumulation tied to instance ownership ( over 4 days and counting)

First I am not certain "Bug reports" is the proper spot but I didn't think POE1 access discussion forum was either. If their is a more appropriate spot please let me know.

I’m reporting a persistent and reproducible latency issue in Path of Exile 2 that has made the game effectively unplayable for me for the past 4+ days. I’ve spent significant time validating that this is not a local PC, GPU, or network issue, and the behavior strongly suggests a server-side backend / instance routing problem.

I’m hoping this can be reviewed from the backend side.

Summary of the Issue
In-game latency accumulates into tens of thousands of milliseconds (30k–70k ms).
The connection never disconnects; instead, latency continuously increases.
This occurs independent of region selection (US and EU tested).
Reconnecting, restarting the client, or switching regions does not reliably resolve it.
The issue has persisted consistently for multiple days.

Key Observations
Stability is instance-ownership dependent:
When I join another player’s instance, gameplay can be stable.
When I return to my own instance, latency immediately becomes unstable again.
If other players join me and I remain in a stable instance, latency remains normal.
Trading / party instance transitions are a very strong trigger.
Latency can also begin mid-map without zoning or player action.

This strongly suggests the problem follows backend instance allocation, not the client or network.

Client-Side Metrics (During Latency Spikes)
Using the in-game performance overlay:
CPU: ~4 ms (stable)
GPU: ~15 ms (stable)
Memory / VRAM: stable
Disk I/O: idle
Network: spiking
Server timing: not reported (expected)

This indicates the client is idle and waiting on server-side processing, not struggling locally.

Network Verification (Ruled Out)
I’ve run multiple WinMTR tests to the PoE2 login endpoint during affected sessions:
0% packet loss at the final hop
Stable ~9–10 ms latency to destination
Any packet loss observed occurs only on intermediate backbone hops and does not propagate downstream (ICMP rate limiting only)

This confirms:
No ISP congestion
No routing instability
No packet loss to GGG edge
No transport-layer latency issues

I can provide WinMTR output if needed.

Client Logs
Client logs show clean login handshakes (~16 ms)
No socket errors, retransmission loops, or disconnect behavior
No CPU/GPU stalls or driver-related errors that correlate with the latency events
Logs are available upon request.

Why This Appears Server-Side
The combination of:
Clean network path
Stable client performance
Latency accumulation without disconnect
Strong correlation to instance ownership and party transitions
Persistence across days and regions

…points to server-side simulation stalls, instance worker contention, or shared backend service blocking, rather than a client or ISP issue.

Request

I’m not asking for a workaround — just that this be reviewed from the backend side, particularly around:

Instance ownership / reassignment
Party & trade instance transitions
Shared services used during instance handoff

If there’s any additional data you’d like me to capture (timestamps, character name, region, etc.), I’m happy to provide it.

Thank you for taking the time to review this.

check (enable/disable) upnp and NAT-PMP in your router

Yeah in this scenario uPNP and NAT-PMP wouldnt cause what I am experiencing or describing.
For SnGs however I went ahead and adjusted, still the same issue occurring.

Case in point on the last hop with WINRT I should/would see packet loss if NAT were the issue, I do not.

Being as I know I am not the only one experiencing this, would be nice for GGG to comment something, especially as I followed their pinned post.