diff --git a/config/manifests/gateway/envoyaigateway/gateway.yaml b/config/manifests/gateway/envoyaigateway/gateway.yaml new file mode 100644 index 000000000..1a536411b --- /dev/null +++ b/config/manifests/gateway/envoyaigateway/gateway.yaml @@ -0,0 +1,17 @@ +apiVersion: gateway.networking.k8s.io/v1 +kind: GatewayClass +metadata: + name: envoy-ai-gateway +spec: + controllerName: gateway.envoyproxy.io/gatewayclass-controller +--- +apiVersion: gateway.networking.k8s.io/v1 +kind: Gateway +metadata: + name: inference-gateway +spec: + gatewayClassName: envoy-ai-gateway + listeners: + - name: http + protocol: HTTP + port: 80 diff --git a/config/manifests/gateway/envoyaigateway/httproute.yaml b/config/manifests/gateway/envoyaigateway/httproute.yaml new file mode 100644 index 000000000..e30b56140 --- /dev/null +++ b/config/manifests/gateway/envoyaigateway/httproute.yaml @@ -0,0 +1,20 @@ +apiVersion: gateway.networking.k8s.io/v1 +kind: HTTPRoute +metadata: + name: llm-route +spec: + parentRefs: + - group: gateway.networking.k8s.io + kind: Gateway + name: inference-gateway + rules: + - backendRefs: + - group: inference.networking.x-k8s.io + kind: InferencePool + name: vllm-llama3-8b-instruct + matches: + - path: + type: PathPrefix + value: / + timeouts: + request: 300s diff --git a/hack/verify-manifests.sh b/hack/verify-manifests.sh index 70d819bc8..dff4170ff 100755 --- a/hack/verify-manifests.sh +++ b/hack/verify-manifests.sh @@ -38,6 +38,7 @@ main() { cp ${SCRIPT_ROOT}/config/crd/bases/* "${TEMP_DIR}/" # Download external CRDs for validation + fetch_crds "https://raw.githubusercontent.com/kubernetes-sigs/gateway-api/refs/tags/${GATEWAY_API_VERSION}/config/crd/standard/gateway.networking.k8s.io_gatewayclasses.yaml" fetch_crds "https://raw.githubusercontent.com/kubernetes-sigs/gateway-api/refs/tags/${GATEWAY_API_VERSION}/config/crd/standard/gateway.networking.k8s.io_gateways.yaml" fetch_crds "https://raw.githubusercontent.com/kubernetes-sigs/gateway-api/refs/tags/${GATEWAY_API_VERSION}/config/crd/standard/gateway.networking.k8s.io_httproutes.yaml" fetch_crds "https://raw.githubusercontent.com/GoogleCloudPlatform/gke-gateway-api/refs/tags/${GKE_GATEWAY_API_VERSION}/config/crd/networking.gke.io_gcpbackendpolicies.yaml"