feat: normalize dynamo namespace computation (#5231)

Signed-off-by: Hannah Zhang <hannahz@nvidia.com>

examples/basics/kubernetes/Distributed_Inference/agg_router.yaml

@@ -8,7 +8,6 @@ metadata:
 spec:
   services:
     Frontend:
-      dynamoNamespace: vllm-agg-router
       componentType: frontend
       replicas: 1
       extraPodSpec:
@@ -19,7 +18,6 @@ spec:
           value: kv
     VllmDecodeWorker:
       envFromSecret: hf-token-secret
-      dynamoNamespace: vllm-agg-router
       componentType: worker
       replicas: 4
       resources:

examples/basics/kubernetes/Distributed_Inference/disagg_router.yaml

@@ -8,7 +8,6 @@ metadata:
 spec:
   services:
     Frontend:
-      dynamoNamespace: vllm-v1-disagg-router
       componentType: frontend
       replicas: 1
       extraPodSpec:
@@ -18,7 +17,6 @@ spec:
         - name: DYN_ROUTER_MODE
           value: kv
     VllmDecodeWorker:
-      dynamoNamespace: vllm-v1-disagg-router
       envFromSecret: hf-token-secret
       componentType: worker
       replicas: 1
@@ -46,7 +44,6 @@ spec:
           args:
             - python3 -m dynamo.vllm --model meta-llama/Llama-3.1-70B-Instruct -tp 8
     VllmPrefillWorker:
-      dynamoNamespace: vllm-v1-disagg-router
       envFromSecret: hf-token-secret
       componentType: worker
       replicas: 1

examples/custom_backend/hello_world/deploy/hello_world.yaml

@@ -29,7 +29,6 @@ spec:
         periodSeconds: 10
         timeoutSeconds: 2
         failureThreshold: 3
-      dynamoNamespace: hello-world
       componentType: frontend
       replicas: 1
       resources:
@@ -68,7 +67,6 @@ spec:
         periodSeconds: 60
         timeoutSeconds: 30
         failureThreshold: 10
-      dynamoNamespace: hello-world
       componentType: worker
       replicas: 1
       resources:

examples/deployments/GKE/sglang/disagg.yaml

@@ -7,7 +7,6 @@ metadata:
 spec:
   services:
     Frontend:
-      dynamoNamespace: sglang-disagg
       componentType: frontend
       replicas: 1
       extraPodSpec:
@@ -15,7 +14,6 @@ spec:
           image: my-registry/sglang-runtime:my-tag
     decode:
       envFromSecret: hf-token-secret
-      dynamoNamespace: sglang-disagg
       componentType: worker
       subComponentType: decode
       replicas: 1
@@ -38,7 +36,6 @@ spec:
             exec python3 -m dynamo.sglang --model-path Qwen/Qwen3-0.6B --served-model-name Qwen/Qwen3-0.6B --page-size 16 --tp 1 --trust-remote-code --skip-tokenizer-init --disaggregation-mode decode --disaggregation-transfer-backend nixl --disaggregation-bootstrap-port --disaggregation-bootstrap-port "12345" --host "0.0.0.0"
     prefill:
       envFromSecret: hf-token-secret
-      dynamoNamespace: sglang-disagg
       componentType: worker
       subComponentType: prefill
       replicas: 1

examples/deployments/GKE/vllm/disagg.yaml

@@ -8,14 +8,12 @@ metadata:
 spec:
   services:
     Frontend:
-      dynamoNamespace: vllm-disagg
       componentType: frontend
       replicas: 1
       extraPodSpec:
         mainContainer:
           image: my-registry/vllm-runtime:my-tag
     VllmDecodeWorker:
-      dynamoNamespace: vllm-disagg
       envFromSecret: hf-token-secret
       componentType: worker
       subComponentType: decode
@@ -39,7 +37,6 @@ spec:
             /sbin/ldconfig
             python3 -m dynamo.vllm --model Qwen/Qwen3-0.6B
     VllmPrefillWorker:
-      dynamoNamespace: vllm-disagg
       envFromSecret: hf-token-secret
       componentType: worker
       subComponentType: prefill

examples/multimodal/deploy/agg_llava.yaml

@@ -9,7 +9,6 @@ spec:
   backendFramework: vllm
   services:
     Frontend:
-      dynamoNamespace: agg-llava
       componentType: frontend
       replicas: 1
       extraPodSpec:
@@ -17,7 +16,6 @@ spec:
           image: my-registry/vllm-runtime:my-tag
     EncodeWorker:
       envFromSecret: hf-token-secret
-      dynamoNamespace: agg-llava
       componentType: worker
       replicas: 1
       resources:
@@ -34,7 +32,6 @@ spec:
             - python3 components/encode_worker.py --model llava-hf/llava-1.5-7b-hf
     VLMWorker:
       envFromSecret: hf-token-secret
-      dynamoNamespace: agg-llava
       componentType: worker
       replicas: 1
       resources:
@@ -51,7 +48,6 @@ spec:
             - python3 components/worker.py --model llava-hf/llava-1.5-7b-hf --worker-type prefill
     Processor:
       envFromSecret: hf-token-secret
-      dynamoNamespace: agg-llava
       componentType: worker
       replicas: 1
       resources:

examples/multimodal/deploy/agg_qwen.yaml

@@ -9,7 +9,6 @@ spec:
   backendFramework: vllm
   services:
     Frontend:
-      dynamoNamespace: agg-qwen
       componentType: frontend
       replicas: 1
       extraPodSpec:
@@ -17,7 +16,6 @@ spec:
           image: my-registry/vllm-runtime:my-tag
     EncodeWorker:
       envFromSecret: hf-token-secret
-      dynamoNamespace: agg-qwen
       componentType: worker
       replicas: 1
       resources:
@@ -34,7 +32,6 @@ spec:
             - python3 components/encode_worker.py --model Qwen/Qwen2.5-VL-7B-Instruct
     VLMWorker:
       envFromSecret: hf-token-secret
-      dynamoNamespace: agg-qwen
       componentType: worker
       replicas: 1
       resources:
@@ -51,7 +48,6 @@ spec:
             - python3 components/worker.py --model Qwen/Qwen2.5-VL-7B-Instruct --worker-type prefill
     Processor:
       envFromSecret: hf-token-secret
-      dynamoNamespace: agg-qwen
       componentType: worker
       replicas: 1
       resources:

recipes/deepseek-r1/sglang/disagg-16gpu/deploy.yaml

@@ -14,7 +14,6 @@ spec:
       create: false
   services:
     Frontend:
-      dynamoNamespace: sgl-dsr1-16gpu
       componentType: frontend
       replicas: 1
       volumeMounts:
@@ -24,7 +23,6 @@ spec:
         mainContainer:
           image: my-registry/sglang-runtime:my-tag
     decode:
-      dynamoNamespace: sgl-dsr1-16gpu
       componentType: worker
       subComponentType: decode
       replicas: 1
@@ -70,7 +68,6 @@ spec:
             - 0.0.0.0
             - --prefill-round-robin-balance
     prefill:
-      dynamoNamespace: sgl-dsr1-16gpu
       componentType: worker
       subComponentType: prefill
       replicas: 1

recipes/deepseek-r1/sglang/disagg-8gpu/deploy.yaml

@@ -14,7 +14,6 @@ spec:
       create: false
   services:
     Frontend:
-      dynamoNamespace: sgl-dsr1-8gpu
       componentType: frontend
       replicas: 1
       volumeMounts:
@@ -24,7 +23,6 @@ spec:
         mainContainer:
           image: my-registry/sglang-runtime:my-tag
     decode:
-      dynamoNamespace: sgl-dsr1-8gpu
       componentType: worker
       subComponentType: decode
       replicas: 1
@@ -67,7 +65,6 @@ spec:
             - 0.0.0.0
             - --prefill-round-robin-balance
     prefill:
-      dynamoNamespace: sgl-dsr1-8gpu
       componentType: worker
       subComponentType: prefill
       replicas: 1

recipes/deepseek-r1/trtllm/disagg/wide_ep/gb200/deploy.yaml

@@ -122,7 +122,6 @@ spec:
   backendFramework: trtllm
   services:
     Frontend:
-      dynamoNamespace: trtllm-disagg-multinode
       componentType: frontend
       replicas: 1
       extraPodSpec:
@@ -137,7 +136,6 @@ spec:
           - /bin/sh
           - -c
     prefill:
-      dynamoNamespace: trtllm-disagg-multinode
       componentType: worker
       subComponentType: prefill
       replicas: 1
@@ -195,7 +193,6 @@ spec:
             configMap:
               name: prefill-config
     decode:
-      dynamoNamespace: trtllm-disagg-multinode
       componentType: worker
       subComponentType: decode
       replicas: 1

recipes/deepseek-r1/vllm/disagg/deploy_hopper_16gpu.yaml

@@ -12,7 +12,6 @@ spec:
       create: false
   services:
     Frontend:
-      dynamoNamespace: vllm-dsr1
       componentType: frontend
       replicas: 1
       extraPodSpec:
@@ -26,7 +25,6 @@ spec:
             failureThreshold: 60
           image: nvcr.io/nvidia/ai-dynamo/vllm-runtime:my-tag
     decode:
-      dynamoNamespace: vllm-dsr1
       componentType: worker
       subComponentType: decode
       replicas: 1
@@ -92,7 +90,6 @@ spec:
                 --max-num-seqs 512 \
                 --compilation_config '{"pass_config":{"enable_fusion":true,"enable_attn_fusion":true,"enable_noop":true},"custom_ops":["+rms_norm"],"cudagraph_mode":"FULL_DECODE_ONLY"}'
     prefill:
-      dynamoNamespace: vllm-dsr1
       componentType: worker
       subComponentType: prefill
       replicas: 1

recipes/gpt-oss-120b/trtllm/agg/deploy.yaml

@@ -28,7 +28,6 @@ spec:
   services:
     Frontend:
       componentType: frontend
-      dynamoNamespace: gpt-oss-agg
       extraPodSpec:
         affinity:
           podAntiAffinity:
@@ -50,7 +49,6 @@ spec:
       replicas: 1
     TrtllmWorker:
       componentType: main
-      dynamoNamespace: gpt-oss-agg
       envFromSecret: hf-token-secret
       volumeMounts:
         - name: model-cache

recipes/llama-3-70b/vllm/agg/deploy.yaml

@@ -12,7 +12,6 @@ spec:
   services:
     Frontend:
       componentType: frontend
-      dynamoNamespace: llama3-70b-agg
       volumeMounts:
         - name: model-cache
           mountPoint: /opt/models
@@ -26,7 +25,6 @@ spec:
       replicas: 1
     VllmPrefillWorker:
       componentType: worker
-      dynamoNamespace: llama3-70b-agg
       envFromSecret: hf-token-secret
       volumeMounts:
         - name: model-cache

recipes/llama-3-70b/vllm/disagg-multi-node/deploy.yaml

@@ -12,7 +12,6 @@ spec:
   services:
     Frontend:
       componentType: frontend
-      dynamoNamespace: llama3-70b-disagg-mn
       volumeMounts:
         - name: model-cache
           mountPoint: /opt/models
@@ -27,7 +26,6 @@ spec:
     VllmPrefillWorker:
       componentType: worker
       subComponentType: prefill
-      dynamoNamespace: llama3-70b-disagg-mn
       envFromSecret: hf-token-secret
       volumeMounts:
         - name: model-cache
@@ -59,7 +57,6 @@ spec:
     VllmDecodeWorker:
       componentType: worker
       subComponentType: decode
-      dynamoNamespace: llama3-70b-disagg-mn
       envFromSecret: hf-token-secret
       volumeMounts:
         - name: model-cache

recipes/llama-3-70b/vllm/disagg-single-node/deploy.yaml

@@ -12,7 +12,6 @@ spec:
   services:
     Frontend:
       componentType: frontend
-      dynamoNamespace: llama3-70b-disagg-sn
       volumeMounts:
         - name: model-cache
           mountPoint: /opt/models
@@ -27,7 +26,6 @@ spec:
     VllmPrefillWorker:
       componentType: worker
       subComponentType: prefill
-      dynamoNamespace: llama3-70b-disagg-sn
       envFromSecret: hf-token-secret
       volumeMounts:
         - name: model-cache
@@ -71,7 +69,6 @@ spec:
     VllmDecodeWorker:
       componentType: worker
       subComponentType: decode
-      dynamoNamespace: llama3-70b-disagg-sn
       envFromSecret: hf-token-secret
       volumeMounts:
         - name: model-cache

recipes/qwen3-235b-a22b-fp8/trtllm/agg/deploy.yaml

@@ -41,7 +41,6 @@ spec:
   services:
     Frontend:
       componentType: frontend
-      dynamoNamespace: qwen3-235b-a22b-agg
       replicas: 1
       extraPodSpec:
         affinity:
@@ -63,7 +62,6 @@ spec:
             - -c
     TrtllmWorker:
       componentType: main
-      dynamoNamespace: qwen3-235b-a22b-agg
       envFromSecret: hf-token-secret
       sharedMemory:
         size: 256Gi

recipes/qwen3-235b-a22b-fp8/trtllm/disagg/deploy.yaml

@@ -70,7 +70,6 @@ spec:
   services:
     Frontend:
       componentType: frontend
-      dynamoNamespace: qwen3-235b-a22b-disagg
       replicas: 1
       extraPodSpec:
         affinity:
@@ -93,7 +92,6 @@ spec:
     TRTLLMPrefillWorker:
       componentType: worker
       subComponentType: prefill
-      dynamoNamespace: qwen3-235b-a22b-disagg
       envFromSecret: hf-token-secret
       replicas: 6
       resources:
@@ -145,7 +143,6 @@ spec:
     TRTLLMDecodeWorker:
       componentType: worker
       subComponentType: decode
-      dynamoNamespace: qwen3-235b-a22b-disagg
       envFromSecret: hf-token-secret
       replicas: 1
       resources:

recipes/qwen3-32b-fp8/trtllm/agg/deploy.yaml

@@ -44,7 +44,6 @@ spec:
   services:
     Frontend:
       componentType: frontend
-      dynamoNamespace: qwen3-32b-fp8-agg
       extraPodSpec:
         affinity:
           podAntiAffinity:
@@ -66,7 +65,6 @@ spec:
       replicas: 1
     TrtllmWorker:
       componentType: main
-      dynamoNamespace: qwen3-32b-fp8-agg
       envFromSecret: hf-token-secret
       volumeMounts:
         - name: model-cache

recipes/qwen3-32b-fp8/trtllm/disagg/deploy.yaml

@@ -201,7 +201,6 @@ spec:
   services:
     Frontend:
       componentType: frontend
-      dynamoNamespace: qwen3-32b-fp8-disagg
       extraPodSpec:
         affinity:
           podAntiAffinity:
@@ -224,7 +223,6 @@ spec:
     TrtllmPrefillWorker:
       componentType: worker
       subComponentType: prefill
-      dynamoNamespace: qwen3-32b-fp8-disagg
       envFromSecret: hf-token-secret
       volumeMounts:
         - name: model-cache
@@ -285,7 +283,6 @@ spec:
     TrtllmDecodeWorker:
       componentType: worker
       subComponentType: decode
-      dynamoNamespace: qwen3-32b-fp8-disagg
       envFromSecret: hf-token-secret
       volumeMounts:
         - name: model-cache

recipes/qwen3-32b/vllm/agg-round-robin/deploy.yaml

@@ -13,7 +13,6 @@ spec:
   services:
     Frontend:
       componentType: frontend
-      dynamoNamespace: agg-8xtp2
       envs:
         - name: HF_HOME
           value: /home/dynamo/.cache/huggingface
@@ -36,7 +35,6 @@ spec:
       subComponentType: null
     VllmDecodeWorker:
       componentType: worker
-      dynamoNamespace: agg-8xtp2
       envFromSecret: hf-token-secret
       volumeMounts:
       - name: model-cache