Using HTTP/2 for load balancing with Ingress

This page shows how to use Kubernetes Ingress and Service objects to configure an an external HTTP(S) load balancer to use HTTP/2 for communication with backend services. This feature is available starting with Google Kubernetes Engine version 1.11.2.

Overview

An HTTP(S) load balancer acts as a proxy between your clients and your application. Clients can use HTTP/1.1 or HTTP/2 to communicate with the load balancer proxy. However, the connection from the load balancer proxy to your application uses HTTP/1.1 by default. If your application, running in a Google Kubernetes Engine pod, is capable of receiving HTTP/2 requests, you configure the external load balancer to use HTTP/2 when it forwards requests to your application.

Diagram that shows HTTP/2 from load balancer to pod (click to enlarge)

In this exercise, you create a Deployment, a Service, and an Ingress. You put a cloud.google.com/app-protocols annotation in your Service manifest to specify that the load balancer should use HTTP/2 to communicate with your application. Then you call your service and verify that your application received an HTTP/2 request.

Before you begin

Before you start, make sure you have performed the following tasks:

Set up default gcloud settings using one of the following methods:

  • Using gcloud init, if you want to be walked through setting defaults.
  • Using gcloud config, to individually set your project ID, zone, and region.

Using gcloud init

If you receive the error One of [--zone, --region] must be supplied: Please specify location, complete this section.

  1. Run gcloud init and follow the directions:

    gcloud init

    If you are using SSH on a remote server, use the --console-only flag to prevent the command from launching a browser:

    gcloud init --console-only
  2. Follow the instructions to authorize gcloud to use your Google Cloud account.
  3. Create a new configuration or select an existing one.
  4. Choose a Google Cloud project.
  5. Choose a default Compute Engine zone.

Using gcloud config

  • Set your default project ID: 
    gcloud config set project PROJECT_ID
  • If you are working with zonal clusters, set your default compute zone: 
    gcloud config set compute/zone COMPUTE_ZONE
  • If you are working with regional clusters, set your default compute region: 
    gcloud config set compute/region COMPUTE_REGION
  • Update gcloud to the latest version: 
    gcloud components update

Creating the Deployment

This Deployment manifest declares that you want to run two replicas of the echoheaders web application:

apiVersion: apps/v1
kind: Deployment
metadata:
  name: echoheaders
spec:
  replicas: 2
  selector:
    matchLabels:
      app: echoheaders
  template:
    metadata:
      labels:
        app: echoheaders
    spec:
      containers:
      - name: echoheaders
        image: k8s.gcr.io/echoserver:1.10
        ports:
        - containerPort: 8443

Copy the manifest to a file named my-deployment.yaml, and create the Deployment:

kubectl apply -f my-deployment.yaml

Creating the Service

Here's a manifest for the Service:

apiVersion: v1
kind: Service
metadata:
  annotations:
    cloud.google.com/app-protocols: '{"my-port":"HTTP2"}'
  name: echoheaders
  labels:
    app: echoheaders
spec:
  type: NodePort
  ports:
  - port: 443
    targetPort: 8443
    protocol: TCP
    name: my-port
  selector:
    app: echoheaders

Save the manifest to a file named my-service.yaml, and create the Service:

kubectl apply -f my-service.yaml

View the Service:

kubectl get service echoheaders --output yaml

The output is similar to this:

apiVersion: v1
kind: Service
metadata:
  annotations:
    cloud.google.com/app-protocols: '{"my-port":"HTTP2"}'
    ...
  labels:
    app: echoheaders
  name: echoheaders
  ...
spec:
  clusterIP: 10.39.251.148
  ...
  ports:
  - name: my-port
    nodePort: 30647
    port: 443
    protocol: TCP
    targetPort: 8443
  selector:
    app: echoheaders
  ...
  type: NodePort
...

For this exercise, here are the important things to note about your Service:

  • The Service has type NodePort. This type is required for Services that are going to be associated with an Ingress.

  • Any Pod that has the label app: echoheaders is a member of the Service. The selector field specifies this.

  • The Service has one port, and the port is named my-port. The cloud.google.com/app-protocols annotation specifies that my-port should use the HTTP/2 protocol.

  • Traffic directed to the service on TCP port 443 is routed to TCP port 8443 in one of the member Pods. The port and targetPort fields specify this.

Creating the Ingress

Here's a manifest for the Ingress:

apiVersion: networking.k8s.io/v1beta1
kind: Ingress
metadata:
  name: echomap
spec:
  backend:
    serviceName: echoheaders
    servicePort: 443

Copy the manifest to a file named my-ingress.yaml, and create the Ingress:

kubectl apply -f my-ingress.yaml

Wait a few minutes for the Kubernetes Ingress controller to configure an HTTP(S) load balancer, and then view the Ingress:

kubectl get ingress echomap --output yaml

The output is similar to this:

kind: Ingress
metadata:
  ...
  name: echomap
  ...
spec:
  backend:
    serviceName: echoheaders
    servicePort: 443
status:
  loadBalancer:
    ingress:
    - ip: 203.0.113.2

For this exercise, here are the important things to note about your Ingress:

  • The IP address for incoming traffic is listed under loadBalancer:ingress.

  • Incoming requests are routed to a Pod that is a member of the echoheaders Service. In this exercise, the member pods have the label app: echoheaders.

  • Requests are routed to the Pod on the target port specified in the echoheaders Service manifest. In this exercise, the Pod target port is 8443.

Verifying that your load balancer supports HTTP/2

gcloud

  1. List your backend services:

    gcloud compute backend-services list

  2. Describe your backend service:

    gcloud beta compute backend-services describe backend-service-name --global

    where backend-service-name is the name of your backend service.

  3. In the output, verify that the protocol is HTTP/2:

    backends:
...
description: '{...,"kubernetes.io/service-port":"443","x-features":["HTTP2"]}'
...
kind: compute#backendService
loadBalancingScheme: EXTERNAL
protocol: HTTP2
...

Console

  1. Go to the Load balancing page in the Google Cloud Console.

    Go to the Load balancing page

  2. Under Name, locate your load balancer.

    Screenshot of HTTP load balancer shown in Google Cloud Console(click to enlarge)

  3. Click the name of your load balancer to view your backend service.

  4. Verify that the Endpoint protocol for your backend service is HTTP/2.

    Screenshot of HTTP/2 backend service shown in Google Cloud Console (click to enlarge)

Calling your service

Wait a few minutes for the load balancer and backend service to be configured. Enter the external IP address of your load balancer in your browser's address bar.

The output shows information about the request from the load balancer to the Pod:

Hostname: echoheaders-7886d5bc68-xnrwj
...
Request Information:
  ...
  method=GET
  real path=/
  query=
  request_version=2
  request_scheme=https
  ...

Request Headers:
  ...
  x-forwarded-for=[YOUR_IP_ADDRESS], 203.0.113.2
  x-forwarded-proto=http
...

For this exercise, here are the important things to note about the preceding output:

  • The line request_version=2 indicates that the request between the load balancer and the Pod used HTTP/2.

  • The line x-forwarded-proto=http indicates that the request between you and the load balancer used HTTP 1.1, not HTTP/2.

What's next?