fix: rm dynamoNamespace from GAIE example (#4313)

Signed-off-by: Anna Tchernych <atchernych@nvidia.com>
Signed-off-by: atchernych <atchernych@nvidia.com>

deploy/inference-gateway/README.md

@@ -111,13 +111,12 @@ cd deploy/inference-gateway
 
 # Export the Dynamo image you have used when deploying your model in Step 3.
 export DYNAMO_IMAGE=<the-dynamo-image-you-have-used-when-deploying-the-model>
-# Export the image tag provided by Dynamo (nvcr.io/nvstaging/ai-dynamo/epp-inference-extension-dynamo:v0.6.0-1) or you can build the Dynamo EPP image by following the commands later in this README.
+# Export the FrontEnd image tag provided by Dynamo (recommended) or build the Dynamo EPP image by following the commands later in this README.
 export EPP_IMAGE=<the-epp-image-you-built>
 ```
 
 ```bash
 helm upgrade --install dynamo-gaie ./helm/dynamo-gaie -n my-model -f ./vllm_agg_qwen.yaml --set-string extension.image=$EPP_IMAGE
-# do not include --set-string extension.image=$EPP_IMAGE to use the default images
 ```
 
 Key configurations include:
@@ -126,7 +125,7 @@ Key configurations include:
 - A service for the inference gateway
 - Required RBAC roles and bindings
 - RBAC permissions
-- values-dynamo-epp.yaml sets epp.dynamo.namespace=vllm-agg for the bundled example. Point it at your actual Dynamo namespace by editing that file or adding --set epp.dynamo.namespace=<namespace> (and likewise for epp.dynamo.component, epp.dynamo.kvBlockSize if they differ).
+- dynamoGraphDeploymentName - the name of the Dynamo Graph where your model is deployed.
 
 
 **Configuration**
@@ -147,7 +146,7 @@ You can configure the plugin by setting environment vars in your [values-dynamo-
 
 Dynamo provides a custom routing plugin `pkg/epp/scheduling/plugins/dynamo_kv_scorer/plugin.go` to perform efficient kv routing.
 The Dynamo router is built as a static library, the EPP router will call to provide fast inference.
-You can either use the image `nvcr.io/nvstaging/ai-dynamo/epp-inference-extension-dynamo:v0.6.0-1` for the EPP_IMAGE in the Helm deployment command and proceed to the step 2 or you can build the image yourself following the steps below.
+You can either use the special FrontEnd image for the EPP_IMAGE in the Helm deployment command and proceed to the step 2 or you can build the image yourself following the steps below.
 
 ##### 1. Build the custom EPP image #####
 

deploy/inference-gateway/helm/dynamo-gaie/templates/_helpers.tpl

@@ -65,6 +65,21 @@ app.kubernetes.io/name: {{ include "dynamo-gaie.name" . }}
 app.kubernetes.io/instance: {{ .Release.Name }}
 {{- end }}
 
+{{/*
+Resolve the fully qualified Dynamo namespace as "<release namespace>-<dynamoGraphDeploymentName>"
+*/}}
+{{- define "dynamo-gaie.dynamoNamespace" -}}
+{{- $dgdName := (.Values.dynamoGraphDeploymentName | default "") | trim -}}
+{{- if not $dgdName }}
+{{- fail "set dynamoGraphDeploymentName to derive the Dynamo namespace" }}
+{{- end }}
+{{- $releaseNamespace := (.Release.Namespace | default "") | trim -}}
+{{- if not $releaseNamespace }}
+{{- fail "Release.Namespace must be set to derive the Dynamo namespace" }}
+{{- end }}
+{{- printf "%s-%s" $releaseNamespace $dgdName }}
+{{- end }}
+
 {{/*
 Create the name of the service account to use
 */}}

deploy/inference-gateway/helm/dynamo-gaie/templates/dynamo-epp.yaml

@@ -16,8 +16,8 @@
 {{- $platformNs   := default .Release.Namespace .Values.platformNamespace -}}
 {{- $platformName := default "dynamo-platform" .Values.platformReleaseName -}}
 {{- $useDynamo    := default false .Values.epp.useDynamo -}}
-{{- $dynNsAll     := default .Values.dynamoNamespace .Values.epp.dynamo.namespace -}}
-{{- $ns           := ternary (required "set epp.dynamo.namespace (or top-level dynamoNamespace) when epp.useDynamo=true" $dynNsAll) "" $useDynamo -}}
+{{- $resolvedDynNs := (include "dynamo-gaie.dynamoNamespace" .) | trim -}}
+{{- $ns           := ternary (required "set dynamoGraphDeploymentName when epp.useDynamo=true" $resolvedDynNs) "" $useDynamo -}}
 {{- $kv           := default "16" .Values.epp.dynamo.kvBlockSize -}}
 {{- $std          := .Values.extension.standardImage -}}
 {{- $dyn          := .Values.extension.dynamoImage -}}
@@ -91,7 +91,7 @@ spec:
             value: "{{ $platformName }}-etcd.{{ $platformNs }}:2379"
           - name: NATS_SERVER
             value: "nats://{{ $platformName }}-nats.{{ $platformNs }}:4222"
-          - name: DYN_NAMESPACE
+          - name: DYNAMO_NAMESPACE
             value: "{{ $ns }}"
           - name: DYNAMO_KV_BLOCK_SIZE
             value: "{{ $kv }}"

deploy/inference-gateway/helm/dynamo-gaie/templates/inference-pool.yaml

@@ -21,7 +21,7 @@ spec:
   targetPortNumber: {{ .Values.inferencePool.port }}
   selector:
     nvidia.com/dynamo-component: Frontend
-    nvidia.com/dynamo-namespace: {{ .Values.dynamoNamespace }}
+    nvidia.com/dynamo-namespace: {{ include "dynamo-gaie.dynamoNamespace" . }}
   extensionRef:
     failureMode: FailOpen
     group: ""

deploy/inference-gateway/helm/dynamo-gaie/values.yaml

@@ -17,8 +17,8 @@
 # This is a YAML-formatted file.
 # Declare variables to be passed into your templates.
 
-# This is the Dynamo namespace where the dynamo model is deployed
-dynamoNamespace: "vllm-agg"
+# Name of the DynamoGraphDeployment (DGD) that backs this gateway integration
+dynamoGraphDeploymentName: "vllm-agg"
 
 # This is the port on which the model is exposed
 model:
@@ -72,7 +72,6 @@ epp:
   # Dynamo-specific settings (only used when useDynamo: true)
   configFile: "/etc/epp/epp-config-dynamo.yaml"
   dynamo:
-    namespace: "vllm-agg" # Required when useDynamo: true.
     kvBlockSize: "16"
 
 # Platform configuration (for Dynamo mode)

deploy/inference-gateway/vllm_agg_qwen.yaml

@@ -5,8 +5,8 @@
 # This is a YAML-formatted file.
 # Declare variables to be passed into your templates.
 
-# This is the Dynamo namespace where the dynamo model is deployed
-dynamoNamespace: "my-model-vllm-agg"
+# Name of the DynamoGraphDeployment (DGD) backing this model deployment.
+dynamoGraphDeploymentName: "vllm-agg"
 
 # This is the port on which the model is exposed
 model:

examples/backends/vllm/deploy/agg.yaml

@@ -13,7 +13,7 @@ spec:
       replicas: 1
       extraPodSpec:
         mainContainer:
-          image: gitlab-master.nvidia.com:5005/dl/ai-dynamo/dynamo/dynamo:vllm-1029
+          image: nvcr.io/nvidia/ai-dynamo/vllm-runtime:my-tag
     VllmDecodeWorker:
       envFromSecret: hf-token-secret
       dynamoNamespace: vllm-agg
@@ -24,7 +24,7 @@ spec:
           gpu: "1"
       extraPodSpec:
         mainContainer:
-          image: gitlab-master.nvidia.com:5005/dl/ai-dynamo/dynamo/dynamo:vllm-1029
+          image: nvcr.io/nvidia/ai-dynamo/vllm-runtime:my-tag
           workingDir: /workspace/examples/backends/vllm
           command:
             - python3