fix: Add --is-decode-worker flag to vllm examples (#5899)

Signed-off-by: Sean Choi <sechoi@nvidia.com>

examples/backends/vllm/launch/disagg.sh

@@ -10,7 +10,7 @@ python -m dynamo.frontend &
 
 # --enforce-eager is added for quick deployment. for production use, need to remove this flag
  DYN_SYSTEM_PORT=${DYN_SYSTEM_PORT1:-8081} \
- CUDA_VISIBLE_DEVICES=0 python3 -m dynamo.vllm --model Qwen/Qwen3-0.6B --enforce-eager &
+ CUDA_VISIBLE_DEVICES=0 python3 -m dynamo.vllm --model Qwen/Qwen3-0.6B --enforce-eager --is-decode-worker &
 
 DYN_SYSTEM_PORT=${DYN_SYSTEM_PORT2:-8082} \
 DYN_VLLM_KV_EVENT_PORT=20081 \

examples/backends/vllm/launch/disagg_router.sh

@@ -24,14 +24,14 @@ CUDA_VISIBLE_DEVICES=0 python3 -m dynamo.vllm \
     --model $MODEL \
     --block-size $BLOCK_SIZE \
     --enforce-eager \
-    --kv-events-config '{"publisher":"zmq","topic":"kv-events","endpoint":"tcp://*:20080","enable_kv_cache_events":true}'&
+    --is-decode-worker &
 
 VLLM_NIXL_SIDE_CHANNEL_PORT=20097 \
 CUDA_VISIBLE_DEVICES=1 python3 -m dynamo.vllm \
     --model $MODEL \
     --block-size $BLOCK_SIZE \
     --enforce-eager \
-    --kv-events-config '{"publisher":"zmq","topic":"kv-events","endpoint":"tcp://*:20081","enable_kv_cache_events":true}' &
+    --is-decode-worker &
 
 # two prefill workers
 # When registered with --is-prefill-worker, these workers are automatically detected

examples/backends/vllm/launch/disagg_router_gaudi.sh

@@ -33,7 +33,7 @@ HABANA_VISIBLE_DEVICES=0 python3 -m dynamo.vllm \
     --block-size $BLOCK_SIZE \
     --kv-transfer-config "{\"kv_connector\": \"NixlConnector\", \"kv_role\": \"kv_both\", \"kv_buffer_device\": \"${NIXL_BUFFER_DEVICE}\", \"kv_connector_extra_config\": {\"backends\": [\"${VLLM_NIXL_BACKEND}\"]}}" \
     --connector none \
-    --kv-events-config '{"publisher":"zmq","topic":"kv-events","endpoint":"tcp://*:5556", "enable_kv_cache_events":true}' &
+    --is-decode-worker &
 
 VLLM_NIXL_SIDE_CHANNEL_PORT=20097 \
 HABANA_VISIBLE_DEVICES=1 python3 -m dynamo.vllm \
@@ -41,7 +41,7 @@ HABANA_VISIBLE_DEVICES=1 python3 -m dynamo.vllm \
     --block-size $BLOCK_SIZE \
     --kv-transfer-config "{\"kv_connector\": \"NixlConnector\", \"kv_role\": \"kv_both\", \"kv_buffer_device\": \"${NIXL_BUFFER_DEVICE}\", \"kv_connector_extra_config\": {\"backends\": [\"${VLLM_NIXL_BACKEND}\"]}}" \
     --connector none \
-    --kv-events-config '{"publisher":"zmq","topic":"kv-events","endpoint":"tcp://*:5557", "enable_kv_cache_events":true}' &
+    --is-decode-worker &
 
 # two prefill workers
 # When registered with --is-prefill-worker, these workers are automatically detected

examples/backends/vllm/launch/disagg_same_gpu.sh

@@ -55,6 +55,7 @@ CUDA_VISIBLE_DEVICES=0 \
 python3 -m dynamo.vllm \
   --model Qwen/Qwen3-0.6B \
   --enforce-eager \
+  --is-decode-worker \
   --gpu-memory-utilization ${GPU_MEM_FRACTION} &
 DECODE_PID=$!