Game version: 1.8.0.13 (Steam)

I am experiencing intermittent 999 ms latency spikes in Arma Reforger, sometimes ending with REPLICATION_STALLED.

The issue occurs on both vanilla / unmodded servers and modded community servers, so Workshop mods are not required to reproduce it.

The issue can also occur when the server is not full. Reaching maximum player capacity is NOT required to reproduce it.

However, I have noticed a strong correlation with server population: when a server reaches or gets very close to its configured maximum capacity, the 999 ms spikes appear to become significantly more frequent.

Examples observed:

39/40 players — issue can occur, but usually less frequently
40/40 players — 999 ms spikes become noticeably more frequent

127/128 players — issue can occur, but usually less frequently
128/128 players — 999 ms spikes become noticeably more frequent

One affected client log contained:

RPL (E): Reliable channel's order sequence appears to be stalled. Terminating connection.

followed by:

Kick cause code: group=1 'REPLICATION', reason=6 'STALLED'

tag: REPLICATION_STALLED

I have already performed extensive troubleshooting:

• Tested both Ethernet and Wi-Fi on the affected PC

• Tested the same PC using a completely different ISP / internet connection — the issue still reproduced

• Tested another PC using my normal router/internet connection — that PC behaved normally

• Full Windows TCP/IP, Winsock and network-device reset

• Router factory reset

• Latest network drivers

• Explicit Windows Firewall allow rules tested

• DNS testing

• Different APN settings

• LTE/5G configuration and band testing

• LTE cell locking

• Tested Steam and Arma Reforger running with administrator privileges

• Reproduced the issue on both vanilla and modded servers

The issue therefore appears to follow this PC / Reforger client rather than one specific network adapter, router, ISP, mobile cell or mod set.

Full server population does not appear to be the sole cause, because the issue can also happen below maximum capacity. My observation is that high/full population makes the problem considerably easier to reproduce or increases its frequency.

Suggested reproduction:

1. Join a high-population server close to its player limit.

2. Observe client latency while the server is below maximum capacity.

3. Continue playing as the server reaches maximum capacity.

4. Compare the frequency of 999 ms spikes before and after reaching maximum capacity.

5. Check client/server replication logs during the spikes.

6. Let the population fall below the maximum again and compare the behavior.

Expected:

Client networking and replication should remain stable regardless of server population.

Actual:

Intermittent 999 ms spikes occur even below maximum capacity, but appear considerably more frequently at high/full population. In some cases the connection eventually terminates with REPLICATION_STALLED.

I can provide complete client logs, screenshots and timestamps from affected sessions if required.

Multiple separate sessions contain the same Reliable channel's order sequence appears to be stalled followed by REPLICATION_STALLED. In one affected session, this occurred four times within approximately one hour.