feat(gateway): add provider-agnostic setup-gateway target (#305)

Signed-off-by: Suraj Deshmukh <suraj.deshmukh@microsoft.com>

Author: surajssd

Makefile

@@ -1,6 +1,6 @@
 .PHONY: install dev dev-frontend dev-backend build compile lint test clean help providers-test verify-versions test-verify-versions
 .PHONY: controller-build controller-docker-build controller-install controller-deploy controller-generate generate-deploy-manifests
-.PHONY: model-downloader-docker-build
+.PHONY: model-downloader-docker-build setup-gateway cleanup-gateway
 
 # Controller image
 CONTROLLER_IMG ?= ghcr.io/kaito-project/airunway/controller:latest
@@ -59,6 +59,10 @@ help:
 	@echo "Provider Targets:"
 	@echo "  providers-test         Run all provider tests"
 	@echo ""
+	@echo "Cluster Setup Targets:"
+	@echo "  setup-gateway          Install Gateway API CRDs, Istio, BBR, and the inference Gateway"
+	@echo "  cleanup-gateway        Remove the inference Gateway and BBR (CRDs/Istio left intact)"
+	@echo ""
 	@echo "Image Build Variables:"
 	@echo "  PLATFORM=<platform>    Target platform for image builds (default: linux/amd64)"
 	@echo "  PUSH=true              Push image instead of loading it locally (default: false)"
@@ -200,6 +204,57 @@ model-downloader-docker-build:
 	docker buildx build --platform $(PLATFORM) $(IMAGE_OUTPUT_FLAG) -f images/model-downloader/Dockerfile -t $(MODEL_DOWNLOADER_IMG) images/model-downloader
 	@echo "✅ Model downloader image built: $(MODEL_DOWNLOADER_IMG) ($(PLATFORM), $(if $(PUSH_ENABLED),pushed,loaded locally))"
 
+# ==================== Cluster Setup Targets ====================
+
+# Provider-agnostic inference-gateway bootstrap: installs Gateway API CRDs,
+# Istio (with the Gateway API Inference Extension enabled), the Body-Based
+# Router, and an `inference-gateway` Gateway resource. Required before
+# deploying any provider that routes through the inference gateway. Versions
+# are pinned in versions.env (GATEWAY_API_VERSION, ISTIO_VERSION, GAIE_VERSION).
+GATEWAY_NAMESPACE ?= default
+GATEWAY_NAME ?= inference-gateway
+GATEWAY_API_URL := https://github.com/kubernetes-sigs/gateway-api/releases/download/$(GATEWAY_API_VERSION)/standard-install.yaml
+GATEWAY_MANIFEST := hack/inference-gateway.yaml
+BBR_CHART := oci://registry.k8s.io/gateway-api-inference-extension/charts/body-based-routing
+GAIE_MANIFEST_URL := https://github.com/kubernetes-sigs/gateway-api-inference-extension/releases/download/$(GAIE_VERSION)/manifests.yaml
+
+setup-gateway: verify-versions
+	@command -v istioctl >/dev/null 2>&1 || { echo "❌ istioctl not found on PATH. Install Istio $(ISTIO_VERSION): https://istio.io/latest/docs/setup/getting-started/"; exit 1; }
+	@echo "Installing Gateway API CRDs $(GATEWAY_API_VERSION)..."
+	kubectl apply -f $(GATEWAY_API_URL)
+	@echo "Installing Gateway API Inference Extension (GAIE) CRDs $(GAIE_VERSION)..."
+	kubectl apply -f $(GAIE_MANIFEST_URL)
+	@echo "Installing Istio $(ISTIO_VERSION) (inference extension enabled)..."
+	istioctl install --skip-confirmation \
+		--set profile=minimal \
+		--set tag=$(ISTIO_VERSION) \
+		--set values.pilot.env.ENABLE_GATEWAY_API_INFERENCE_EXTENSION=true
+	@echo "Installing Body-Based Router (BBR) $(GAIE_VERSION) into namespace $(GATEWAY_NAMESPACE)..."
+	helm upgrade -i body-based-router \
+		--namespace $(GATEWAY_NAMESPACE) --create-namespace \
+		--set provider.name=istio \
+		--version "$(GAIE_VERSION)" \
+		--wait \
+		$(BBR_CHART)
+	@echo "Creating Gateway resource $(GATEWAY_NAME) in namespace $(GATEWAY_NAMESPACE)..."
+	@command -v envsubst >/dev/null 2>&1 || { echo "❌ envsubst not found on PATH (provided by gettext)."; exit 1; }
+	GATEWAY_NAME=$(GATEWAY_NAME) GATEWAY_NAMESPACE=$(GATEWAY_NAMESPACE) \
+		envsubst < $(GATEWAY_MANIFEST) | kubectl apply -f -
+	@echo "✅ Inference gateway ready (Istio $(ISTIO_VERSION), gateway/$(GATEWAY_NAME))"
+
+# Tear down the inference Gateway and BBR. Gateway API CRDs and Istio are left
+# intact because they may be shared with other workloads.
+cleanup-gateway:
+	@command -v envsubst >/dev/null 2>&1 || { echo "❌ envsubst not found on PATH (provided by gettext)."; exit 1; }
+	@GATEWAY_NAME=$(GATEWAY_NAME) GATEWAY_NAMESPACE=$(GATEWAY_NAMESPACE) \
+		envsubst < $(GATEWAY_MANIFEST) | kubectl delete -f - --ignore-not-found
+	@helm uninstall body-based-router --namespace $(GATEWAY_NAMESPACE) --ignore-not-found || helm uninstall body-based-router --namespace $(GATEWAY_NAMESPACE) || true
+	@echo "⚠️ Gateway API CRDs, GAIE CRDs, and Istio left intact (may be shared). Remove manually if needed:"
+	@echo "    kubectl delete -f \"$(GATEWAY_API_URL)\" --ignore-not-found"
+	@echo "    kubectl delete -f \"$(GAIE_MANIFEST_URL)\" --ignore-not-found"
+	@echo "    istioctl uninstall --purge -y"
+	@echo "✅ Inference gateway and BBR removed"
+
 # ==================== Version Drift Guard ====================
 
 # Verify all version references are in sync with versions.env.

docs/gateway.md

@@ -74,6 +74,16 @@ AI Runway works with any Gateway API implementation that supports the [Inference
 
 ## Setup
 
+> [!TIP]
+> **Istio shortcut:** `make setup-gateway` (from the repo root) performs the entire manual
+> **Istio** setup below in one shot — it installs the Gateway API CRDs (Step 1), the Gateway
+> API Inference Extension (GAIE) CRDs (Step 2), Istio with the inference extension enabled
+> (Step 3), the `inference-gateway` Gateway resource (Step 4), and the Body-Based Router (see
+> [Body-Based Routing](#body-based-routing-bbr)). The
+> `GATEWAY_API_VERSION`, `ISTIO_VERSION`, and `GAIE_VERSION` it uses are pinned in
+> [`/versions.env`](../versions.env), and `istioctl` must be on your PATH. For other gateway
+> implementations, follow the manual steps below.
+
 ### Step 1: Install Gateway API CRDs
 
 ```bash

hack/inference-gateway.yaml

@@ -0,0 +1,26 @@
+# Provider-agnostic inference gateway used by the root `setup-gateway` target.
+# The name and namespace are filled in by envsubst at apply time
+# (defaults: inference-gateway / default — see GATEWAY_NAME / GATEWAY_NAMESPACE
+# in the root Makefile).
+apiVersion: gateway.networking.k8s.io/v1
+kind: Gateway
+metadata:
+  name: ${GATEWAY_NAME}
+  namespace: ${GATEWAY_NAMESPACE}
+  labels:
+    # Identifies this as the airunway inference gateway. The controller's
+    # auto-selection and the backend status check use this label to
+    # disambiguate when more than one Gateway exists in the cluster (see
+    # controller/internal/gateway/detection.go and gateway_reconciler.go).
+    airunway.ai/inference-gateway: "true"
+spec:
+  gatewayClassName: istio
+  infrastructure:
+    annotations:
+      # Required on AKS with Istio. Azure otherwise probes GET / on port 80,
+      # but the gateway returns 404 there and the public IP can time out.
+      service.beta.kubernetes.io/port_80_health-probe_protocol: tcp
+  listeners:
+    - name: http
+      protocol: HTTP
+      port: 80

providers/dynamo/Makefile

@@ -70,13 +70,14 @@ setup-dynamo: verify-versions
 		curl -fsSL "$(DYNAMO_PRE_DEPLOY_URL)" -o "$$TMPSCRIPT" && \
 		chmod +x "$$TMPSCRIPT" && \
 		bash "$$TMPSCRIPT"
-	@echo "Pre-deployment checks passed. Installing GAIE components..."
-	@kubectl apply -f $(GAIE_MANIFEST_URL)
+	@echo "Pre-deployment checks passed. Setting up the inference gateway (Gateway API CRDs, GAIE CRDs, Istio, BBR, Gateway)..."
+	@$(MAKE) -C ../.. setup-gateway
 	@echo "Installing Dynamo platform v$(DYNAMO_VERSION)..."
 	@helm upgrade -i dynamo-platform "$(DYNAMO_CHART_URL)" \
 			--namespace $(DYNAMO_NAMESPACE) \
 			--create-namespace \
-			--set "dynamo-operator.controllerManager.manager.image.tag=$(DYNAMO_VERSION)"
+			--set "dynamo-operator.controllerManager.manager.image.tag=$(DYNAMO_VERSION)" \
+			--wait
 	@echo "✅ Dynamo platform v$(DYNAMO_VERSION) installed in namespace $(DYNAMO_NAMESPACE)"
 
 ## Uninstall Dynamo platform from the cluster

shared/types/versions.generated.ts

@@ -5,3 +5,5 @@
 export const GAIE_VERSION = "v1.5.0";
 export const DYNAMO_VERSION = "1.1.1";
 export const KAITO_VERSION = "0.10.0";
+export const GATEWAY_API_VERSION = "v1.5.1";
+export const ISTIO_VERSION = "1.30.0";

versions.env

@@ -24,3 +24,11 @@ DYNAMO_VERSION=1.1.1
 
 # KAITO workspace operator Helm chart version (kaito/workspace chart).
 KAITO_VERSION=0.10.0
+
+# Gateway API CRDs (a separate project from GAIE). Used by the root
+# `setup-gateway` Makefile target to install standard-install.yaml.
+GATEWAY_API_VERSION=v1.5.1
+
+# Istio control plane installed by the root `setup-gateway` target
+# (gatewayClassName: istio). Requires a matching local istioctl on PATH.
+ISTIO_VERSION=1.30.0