Configuring Canary Deployment Ingress

This article will help you deploy a Kubernetes cluster and configure Canary Deployment on it using Nginx Ingress Annotations: execute the Canary Deployment script for the echo server and make sure that traffic is distributed according to the configuration file.

1. Create a Kubernetes cluster of the latest version, if not done so already.

2. Connect to the cluster using kubectl.

3. Create a test application:

   1. Create a new namespace for the project:

       
      kubectl create ns echo-production

   2. Create a Kubernetes resource based on the manifest, for example, http-svc:

    
   kubectl apply -f https://raw.githubusercontent.com/kubernetes/ingress-nginx/master/docs/examples/http-svc.yaml -n echo-production

1. Create a manifest file http-svc.ingress with the following contents:

   http-svc.ingress

2. Apply the manifest in the cluster:

    
   kubectl apply -f http-svc.ingress -n echo-production

   As a result, an application will be created and the server will respond to all requests from echo.com.

1. Create a Canary version of the namespace for the application:

    
   kubectl create ns echo-canary

2. Deploy the Canary version of the app:

    
   kubectl apply -f https://raw.githubusercontent.com/kubernetes/ingress-nginx/master/docs/examples/http-svc.yaml -n echo-canary

3. Create a Canary version of the Ingress configuration file:

   http-svc.ingress.canary

4. Apply the manifest in the cluster:

    
   kubectl apply -f http-svc.ingress.canary -n echo-canary

1. Connect to the cluster using Kubernetes Dashboard.

2. Go to Namespaces.

3. Switch the Namespace filter to All.

4. At the bottom of the side menu, select Ingresses.

   A list of all available Ingresses will be displayed.

5. Make sure that http-svc has one IP address specified in the Endpoints column.

6. Check the distribution of requests according to the installed configuration by executing the script count.rb:

   count.rb

    
   ruby count.rb

Example of a successful result:

 
{"echo-production"=>896, "echo-canary"=>104}

A running cluster consumes computing resources and is charged accordingly. If you created it for testing purposes and no longer need it:

• Stop it to use it later.
• Delete it permanently.