fix: set frontend defaults

components/backends/sglang/deploy/agg.yaml

@@ -21,10 +21,6 @@ spec:
       extraPodSpec:
         mainContainer:
           image: my-registry/sglang-runtime:my-tag
-          workingDir: /workspace/components/backends/sglang
-          command: ["sh", "-c"]
-          args:
-            - "python3 -m dynamo.sglang.utils.clear_namespace --namespace sglang-agg && python3 -m dynamo.frontend --http-port=8000"
     SGLangDecodeWorker:
       envFromSecret: hf-token-secret
       dynamoNamespace: sglang-agg

components/backends/sglang/deploy/agg_router.yaml

@@ -21,10 +21,9 @@ spec:
       extraPodSpec:
         mainContainer:
           image: my-registry/sglang-runtime:my-tag
-          workingDir: /workspace/components/backends/sglang
-          command: ["sh", "-c"]
-          args:
-            - "python3 -m dynamo.sglang.utils.clear_namespace --namespace sglang-agg-router && python3 -m dynamo.frontend --http-port=8000  --router-mode kv"
+      envs:
+        - name: DYNAMO_ROUTER_MODE
+          value: kv
     SGLangDecodeWorker:
       envFromSecret: hf-token-secret
       dynamoNamespace: sglang-agg-router

components/backends/sglang/deploy/disagg-multinode.yaml

@@ -23,10 +23,6 @@ spec:
       extraPodSpec:
         mainContainer:
           image: my-registry/sglang-runtime:my-tag
-          workingDir: /workspace/components/backends/sglang
-          command: ["sh", "-c"]
-          args:
-            - "python3 -m dynamo.sglang.utils.clear_namespace --namespace sglang-disagg-multinode && python3 -m dynamo.frontend --http-port=8000"
     decode:
       multinode:
         nodeCount: 2

components/backends/sglang/deploy/disagg.yaml

@@ -21,10 +21,6 @@ spec:
       extraPodSpec:
         mainContainer:
           image: my-registry/sglang-runtime:my-tag
-          workingDir: /workspace/components/backends/sglang
-          command: ["sh", "-c"]
-          args:
-            - "python3 -m dynamo.sglang.utils.clear_namespace --namespace sglang-disagg && python3 -m dynamo.frontend --http-port=8000"
     SGLangDecodeWorker:
       envFromSecret: hf-token-secret
       dynamoNamespace: sglang-disagg

components/backends/sglang/deploy/disagg_planner.yaml

@@ -28,10 +28,6 @@ spec:
       extraPodSpec:
         mainContainer:
           image: nvcr.io/nvidian/nim-llm-dev/sglang-runtime:hzhou-0811-1
-          workingDir: /workspace/components/backends/sglang
-          command: ["sh", "-c"]
-          args:
-            - "python3 -m dynamo.sglang.utils.clear_namespace --namespace sglang-disagg && python3 -m dynamo.frontend --http-port=8000"
     Planner:
       dynamoNamespace: dynamo
       envFromSecret: hf-token-secret

components/backends/trtllm/deploy/agg.yaml

@@ -21,12 +21,6 @@ spec:
       extraPodSpec:
         mainContainer:
           image: nvcr.io/nvidian/nim-llm-dev/trtllm-runtime:dep-233.17
-          workingDir: /workspace/components/backends/trtllm
-          command:
-            - /bin/sh
-            - -c
-          args:
-            - "python3 -m dynamo.frontend --http-port 8000"
     TRTLLMWorker:
       envFromSecret: hf-token-secret
       dynamoNamespace: trtllm-agg

components/backends/trtllm/deploy/agg_router.yaml

@@ -14,12 +14,9 @@ spec:
       extraPodSpec:
         mainContainer:
           image: nvcr.io/nvidian/nim-llm-dev/trtllm-runtime:dep-233.17
-          workingDir: /workspace/components/backends/trtllm
-          command:
-            - /bin/sh
-            - -c
-          args:
-            - "python3 -m dynamo.frontend --http-port 8000 --router-mode kv"
+      envs:
+        - name: DYNAMO_ROUTER_MODE
+          value: kv
     TRTLLMWorker:
       envFromSecret: hf-token-secret
       dynamoNamespace: trtllm-agg-router

components/backends/trtllm/deploy/disagg.yaml

@@ -21,12 +21,6 @@ spec:
       extraPodSpec:
         mainContainer:
           image: nvcr.io/nvidian/nim-llm-dev/trtllm-runtime:dep-233.17
-          workingDir: /workspace/components/backends/trtllm
-          command:
-            - /bin/sh
-            - -c
-          args:
-            - "python3 -m dynamo.frontend --http-port 8000"
     TRTLLMPrefillWorker:
       dynamoNamespace: trtllm-disagg
       envFromSecret: hf-token-secret

components/backends/trtllm/deploy/disagg_router.yaml

@@ -21,12 +21,9 @@ spec:
       extraPodSpec:
         mainContainer:
           image: nvcr.io/nvidian/nim-llm-dev/trtllm-runtime:dep-233.17
-          workingDir: /workspace/components/backends/trtllm
-          command:
-            - /bin/sh
-            - -c
-          args:
-            - "python3 -m dynamo.frontend --http-port 8000 --router-mode kv"
+      envs:
+        - name: DYNAMO_ROUTER_MODE
+          value: kv
     TRTLLMPrefillWorker:
       dynamoNamespace: trtllm-v1-disagg-router
       envFromSecret: hf-token-secret

components/backends/vllm/deploy/agg.yaml

@@ -14,12 +14,6 @@ spec:
       extraPodSpec:
         mainContainer:
           image: nvcr.io/nvidian/nim-llm-dev/vllm-runtime:dep-233.17
-          workingDir: /workspace/components/backends/vllm
-          command:
-            - /bin/sh
-            - -c
-          args:
-            - "python3 -m dynamo.frontend --http-port 8000"
     VllmDecodeWorker:
       envFromSecret: hf-token-secret
       dynamoNamespace: vllm-agg

components/backends/vllm/deploy/agg_router.yaml

@@ -14,12 +14,9 @@ spec:
       extraPodSpec:
         mainContainer:
           image: nvcr.io/nvidian/nim-llm-dev/vllm-runtime:dep-233.17
-          workingDir: /workspace/components/backends/vllm
-          command:
-            - /bin/sh
-            - -c
-          args:
-            - "python3 -m dynamo.frontend --http-port 8000 --router-mode kv"
+      envs:
+        - name: DYNAMO_ROUTER_MODE
+          value: kv
     VllmDecodeWorker:
       envFromSecret: hf-token-secret
       dynamoNamespace: vllm-agg-router

components/backends/vllm/deploy/disagg.yaml

@@ -21,12 +21,6 @@ spec:
       extraPodSpec:
         mainContainer:
           image: nvcr.io/nvidian/nim-llm-dev/vllm-runtime:dep-233.17
-          workingDir: /workspace/components/backends/vllm
-          command:
-            - /bin/sh
-            - -c
-          args:
-            - "python3 -m dynamo.frontend --http-port 8000"
     VllmDecodeWorker:
       dynamoNamespace: vllm-disagg
       envFromSecret: hf-token-secret

components/backends/vllm/deploy/disagg_planner.yaml

@@ -30,12 +30,6 @@ spec:
       extraPodSpec:
         mainContainer:
           image: nvcr.io/nvidian/nim-llm-dev/vllm-runtime:hzhou-0814-02
-          workingDir: /workspace/components/backends/vllm
-          command:
-            - /bin/sh
-            - -c
-          args:
-            - "python3 -m dynamo.frontend --http-port 8000"
     Planner:
       dynamoNamespace: vllm-disagg-planner
       envFromSecret: hf-token-secret

components/backends/vllm/deploy/disagg_router.yaml

@@ -14,12 +14,9 @@ spec:
       extraPodSpec:
         mainContainer:
           image: nvcr.io/nvidian/nim-llm-dev/vllm-runtime:dep-233.17
-          workingDir: /workspace/components/backends/vllm
-          command:
-            - /bin/sh
-            - -c
-          args:
-            - "python3 -m dynamo.frontend --http-port 8000 --router-mode kv"
+      envs:
+        - name: DYNAMO_ROUTER_MODE
+          value: kv
     VllmDecodeWorker:
       dynamoNamespace: vllm-v1-disagg-router
       envFromSecret: hf-token-secret

components/frontend/src/dynamo/frontend/main.py

@@ -83,14 +83,17 @@ def parse_args():
         "--kv-cache-block-size", type=int, help="KV cache block size (u32)."
     )
     parser.add_argument(
-        "--http-port", type=int, default=8080, help="HTTP port for the engine (u16)."
+        "--http-port", 
+        type=int, 
+        default=int(os.environ.get("DYNAMO_HTTP_PORT", "8080")), 
+        help="HTTP port for the engine (u16). Can be set via DYNAMO_HTTP_PORT env var."
     )
     parser.add_argument(
         "--router-mode",
         type=str,
         choices=["round-robin", "random", "kv"],
-        default="round-robin",
-        help="How to route the request",
+        default=os.environ.get("DYNAMO_ROUTER_MODE", "round-robin"),
+        help="How to route the request. Can be set via DYNAMO_ROUTER_MODE env var.",
     )
     parser.add_argument(
         "--kv-overlap-score-weight",

deploy/cloud/operator/internal/dynamo/component_frontend.go

@@ -27,6 +27,10 @@ func (f *FrontendDefaults) GetBaseContainer(context ComponentContext) (corev1.Co
 	// Frontend doesn't need backend-specific config
 	container := f.getCommonContainer(context)
 
+	// Set default command and args
+	container.Command = []string{"python3"}
+	container.Args = []string{"-m", "dynamo.frontend"}
+
 	// Add HTTP port
 	container.Ports = []corev1.ContainerPort{
 		{
@@ -83,6 +87,10 @@ func (f *FrontendDefaults) GetBaseContainer(context ComponentContext) (corev1.Co
 			Name:  commonconsts.EnvDynamoServicePort,
 			Value: fmt.Sprintf("%d", commonconsts.DynamoServicePort),
 		},
+		{
+			Name:  "DYNAMO_HTTP_PORT", // TODO: need to reconcile DYNAMO_PORT and DYNAMO_HTTP_PORT
+			Value: fmt.Sprintf("%d", commonconsts.DynamoServicePort),
+		},
 	}...)
 
 	return container, nil

deploy/cloud/operator/internal/dynamo/graph_test.go

@@ -1283,6 +1283,10 @@ func TestGenerateGrovePodGangSet(t *testing.T) {
 													},
 												},
 												Env: []corev1.EnvVar{
+													{
+														Name:  "DYNAMO_HTTP_PORT",
+														Value: fmt.Sprintf("%d", commonconsts.DynamoServicePort),
+													},
 													{
 														Name:  "DYNAMO_POD_GANG_SET_REPLICAS",
 														Value: "1",
@@ -2020,6 +2024,10 @@ func TestGenerateGrovePodGangSet(t *testing.T) {
 													},
 												},
 												Env: []corev1.EnvVar{
+													{
+														Name:  "DYNAMO_HTTP_PORT",
+														Value: fmt.Sprintf("%d", commonconsts.DynamoServicePort),
+													},
 													{
 														Name:  "DYNAMO_POD_GANG_SET_REPLICAS",
 														Value: "1",
@@ -2769,6 +2777,10 @@ func TestGenerateGrovePodGangSet(t *testing.T) {
 													},
 												},
 												Env: []corev1.EnvVar{
+													{
+														Name:  "DYNAMO_HTTP_PORT",
+														Value: fmt.Sprintf("%d", commonconsts.DynamoServicePort),
+													},
 													{
 														Name:  "DYNAMO_POD_GANG_SET_REPLICAS",
 														Value: "1",
@@ -4211,6 +4223,87 @@ func XTestGenerateGrovePodGangSet_StartsAfterDependencies(t *testing.T) {
 	}
 }
 
+func TestGenerateBasePodSpec_Frontend(t *testing.T) {
+	secretsRetriever := &mockSecretsRetriever{}
+	controllerConfig := controller_common.Config{}
+	dynamoDeployment := &v1alpha1.DynamoGraphDeployment{
+		ObjectMeta: metav1.ObjectMeta{
+			Name:      "test-deployment",
+			Namespace: "default",
+		},
+	}
+
+	tests := []struct {
+		name             string
+		component        *v1alpha1.DynamoComponentDeploymentOverridesSpec
+		backendFramework BackendFramework
+		wantEnvVars      map[string]string
+		wantErr          bool
+	}{
+		{
+			name: "frontend with default command",
+			component: &v1alpha1.DynamoComponentDeploymentOverridesSpec{
+				DynamoComponentDeploymentSharedSpec: v1alpha1.DynamoComponentDeploymentSharedSpec{
+					ComponentType: commonconsts.ComponentTypeFrontend,
+				},
+			},
+			backendFramework: BackendFrameworkVLLM,
+			wantEnvVars: map[string]string{
+				"DYNAMO_HTTP_PORT": fmt.Sprintf("%d", commonconsts.DynamoServicePort),
+			},
+		},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			podSpec, err := GenerateBasePodSpec(
+				tt.component,
+				tt.backendFramework,
+				secretsRetriever,
+				dynamoDeployment.Name,
+				dynamoDeployment.Namespace,
+				RoleMain,
+				1,
+				controllerConfig,
+				commonconsts.MultinodeDeploymentTypeGrove,
+				"test-service",
+			)
+
+			if (err != nil) != tt.wantErr {
+				t.Errorf("GenerateBasePodSpec() error = %v, wantErr %v", err, tt.wantErr)
+				return
+			}
+			if tt.wantErr {
+				return
+			}
+
+			// Check command and args
+			wantCommand := []string{"python3"}
+			wantArgs := []string{"-m", "dynamo.frontend"}
+			if !reflect.DeepEqual(podSpec.Containers[0].Command, wantCommand) {
+				t.Errorf("GenerateBasePodSpec() command = %v, want %v",
+					podSpec.Containers[0].Command, wantCommand)
+			}
+			if !reflect.DeepEqual(podSpec.Containers[0].Args, wantArgs) {
+				t.Errorf("GenerateBasePodSpec() args = %v, want %v",
+					podSpec.Containers[0].Args, wantArgs)
+			}
+
+			// Check environment variables
+			envVars := make(map[string]string)
+			for _, env := range podSpec.Containers[0].Env {
+				envVars[env.Name] = env.Value
+			}
+			for k, v := range tt.wantEnvVars {
+				if envVars[k] != v {
+					t.Errorf("GenerateBasePodSpec() env var %s = %v, want %v",
+						k, envVars[k], v)
+				}
+			}
+		})
+	}
+}
+
 func TestGenerateBasePodSpec_PlannerServiceAccount(t *testing.T) {
 	secretsRetriever := &mockSecretsRetriever{}
 	controllerConfig := controller_common.Config{}