fix: Remove asymmetric --request-plane nats from run_engines.sh script (#5245)

benchmarks/router/README.md

@@ -107,6 +107,9 @@ We also supports running lightweight mock engines that simulate vLLM behavior wi
 In a **new terminal**, launch the Dynamo router using the Python CLI:
 
 ```bash
+# Explicitly set NATS server for KV event publishing
+export NATS_SERVER="${NATS_SERVER:-nats://localhost:4222}"
+
 python -m dynamo.frontend \
     --router-mode kv \
     --router-reset-states \

benchmarks/router/run_engines.sh

@@ -251,7 +251,6 @@ else
                     VLLM_ARGS+=("--is-decode-worker")
                 fi
                 VLLM_ARGS+=("${EXTRA_ARGS[@]}")
-                VLLM_ARGS+=("--request-plane" "nats")
 
                 exec env PYTHONHASHSEED=0 CUDA_VISIBLE_DEVICES=$GPU_DEVICES DYN_VLLM_KV_EVENT_PORT=$((20080 + i)) VLLM_NIXL_SIDE_CHANNEL_PORT=$((20096 + i)) python3 -m dynamo.vllm \
                     "${VLLM_ARGS[@]}"
@@ -270,4 +269,4 @@ fi
 
 echo "All workers started. Press Ctrl+C to stop."
 wait
-echo "All workers completed."
\ No newline at end of file
+echo "All workers completed."