Implementation:Tensorflow Serving Resnet K8s Manifest

Knowledge Sources	TensorFlow Serving
Domains	Kubernetes, Deployment
Last Updated	2026-02-13 17:00 GMT

Overview

Concrete Kubernetes manifest defining a Deployment and LoadBalancer Service for serving a ResNet model on GKE.

Description

resnet_k8s.yaml defines two Kubernetes resources:

• Deployment (resnet-deployment): 3 replicas of the gcr.io/tensorflow-serving/resnet image, each exposing port 8500 (gRPC). Pods are labeled app: resnet-server for service selection.
• Service (resnet-service): LoadBalancer service routing external port 8500 to container port 8500, selecting pods with app: resnet-server label.

This is a pattern document — users should customize the image, replica count, and port configuration for their models.

Usage

Apply with kubectl create -f after pushing the Docker image and creating the cluster. Modify the image URL and model name for your specific model.

Code Reference

Source Location

• Repository: tensorflow/serving
• File: tensorflow_serving/example/resnet_k8s.yaml (L1-49)

Signature

apiVersion: apps/v1
kind: Deployment
metadata:
  name: resnet-deployment
spec:
  selector:
    matchLabels:
      app: resnet-server
  replicas: 3
  template:
    metadata:
      labels:
        app: resnet-server
    spec:
      containers:
      - name: resnet-container
        image: gcr.io/tensorflow-serving/resnet
        ports:
        - containerPort: 8500
---
apiVersion: v1
kind: Service
metadata:
  labels:
    run: resnet-service
  name: resnet-service
spec:
  ports:
  - port: 8500
    targetPort: 8500
  selector:
    app: resnet-server
  type: LoadBalancer

Import

kubectl create -f tensorflow_serving/example/resnet_k8s.yaml

I/O Contract

Inputs

Name	Type	Required	Description
YAML manifest	file	Yes	Kubernetes resource definitions
Docker image in registry	URL	Yes	Container image accessible from cluster
Running cluster	cluster	Yes	Kubernetes cluster with kubectl configured

Outputs

Name	Type	Description
Deployment	Kubernetes resource	3 pods running tensorflow_model_server
Service	Kubernetes resource	LoadBalancer with external IP on port 8500

Usage Examples

Deploy ResNet

# Apply manifest
kubectl create -f tensorflow_serving/example/resnet_k8s.yaml

# Check deployment status
kubectl get deployments
kubectl get pods

# Get external IP
kubectl get services resnet-service
# Wait for EXTERNAL-IP to be assigned

# Scale up
kubectl scale deployment resnet-deployment --replicas=5

# Update image (rolling update)
kubectl set image deployment/resnet-deployment \
    resnet-container=gcr.io/my-project/resnet_serving:v2

Related Pages

Implements Principle

• Principle:Tensorflow_Serving_Kubernetes_Resource_Deployment

Requires Environment

• Environment:Tensorflow_Serving_Kubernetes_Deployment_Environment