4.traffic-splitting

Traffic Splitting

This use case configures traffic splitting for a sample application with two services: coffee-v1-svc and coffee-v2-svc 90% of the coffee application traffic is sent to coffee-v1-svc the remaining 10% to coffee-v2-svc

This use case shows how to enforce JWT authentication at the NGINX Ingress Controller level

Get NGINX Ingress Controller Node IP, HTTP and HTTPS NodePorts

export NIC_IP=`kubectl get pod -l app.kubernetes.io/instance=nic -n nginx-ingress -o json|jq '.items[0].status.hostIP' -r`
export HTTP_PORT=`kubectl get svc nic-nginx-ingress-controller -n nginx-ingress -o jsonpath='{.spec.ports[0].nodePort}'`
export HTTPS_PORT=`kubectl get svc nic-nginx-ingress-controller -n nginx-ingress -o jsonpath='{.spec.ports[1].nodePort}'`

Check NGINX Ingress Controller IP address, HTTP and HTTPS ports

echo -e "NIC address: $NIC_IP\nHTTP port  : $HTTP_PORT\nHTTPS port : $HTTPS_PORT"

cd into the lab directory

cd ~/NGINX-Ingress-Controller-Lab/labs/4.traffic-splitting

Deploy the sample web applications

kubectl apply -f 0.cafe.yaml

Publish the application through NGINX Ingress Controller applying the traffic splitting rule

kubectl apply -f 1.virtual-server.yaml

Check the newly created VirtualServer resource

kubectl get vs -o wide

Output should be similar to

NAME   STATE   HOST               IP    EXTERNALHOSTNAME   PORTS   AGE
cafe   Valid   cafe.example.com                                    3s

Describe the cafe virtualserver

kubectl describe vs cafe

Output should be similar to

Name:         cafe
Namespace:    default
Labels:       <none>
Annotations:  <none>
API Version:  k8s.nginx.org/v1
Kind:         VirtualServer
Metadata:
  Creation Timestamp:  2025-04-03T20:41:07Z
  Generation:          1
  Resource Version:    247959
  UID:                 861ef737-ef95-4a9e-875b-e83a8f9c7f3a
Spec:
  Host:  cafe.example.com
  Routes:
    Path:  /coffee
    Splits:
      Action:
        Pass:  coffee-v1
      Weight:  90
      Action:
        Pass:  coffee-v2
      Weight:  10
  Upstreams:
    Name:     coffee-v1
    Port:     80
    Service:  coffee-v1-svc
    Name:     coffee-v2
    Port:     80
    Service:  coffee-v2-svc
Status:
  Message:  Configuration for default/cafe was added or updated 
  Reason:   AddedOrUpdated
  State:    Valid
Events:
  Type    Reason          Age   From                      Message
  ----    ------          ----  ----                      -------
  Normal  AddedOrUpdated  24s   nginx-ingress-controller  Configuration for default/cafe was added or updated

Test application access

Access the application

curl -H "Host: cafe.example.com" http://$NIC_IP:$HTTP_PORT/coffee

90% of responses will come from coffee-v1-svc and be similar to

Server address: 192.168.36.105:8080
Server name: coffee-v1-c48b96b65-6rlgx
Date: 03/Apr/2025:20:42:14 +0000
URI: /coffee
Request ID: 7e5a8e7143f9f5341b9c19ee60e47b8b

10% of responses will come from coffee-v2-svc and be similar to

Server address: 192.168.36.104:8080
Server name: coffee-v2-685fd9bb65-xpgfl
Date: 03/Apr/2025:20:42:48 +0000
URI: /coffee
Request ID: 77d52bf9dc70bd6df6ee099553835615

Test access using the script provided. It sends 100 requests and shows the traffic split ratio

. ./test.sh

Output should be similar to

Summary of responses:
Coffee v1: 90 times
Coffee v2: 10 times

Delete the lab

kubectl delete -f .