After the script has finished, three the key components will have been deployed.

Aspen Mesh

Aspen Mesh is a hosted solution. The data still needs to be pushed into the system. This happens via an Agent that is deployed. You can view the deployed pod with:

kubectl get pods -n istio-system | grep aspen

Istio

Istio is a Service Mesh that provides a uniform way to secure, connect, and monitor microservices. This is deployed within its own namespace, called istio-system. Aspen Mesh builds on top of Istio.

kubectl get pods -n istio-system

BookInfo Sample Application

As part of the deployment script for Aspen Mesh, the BookInfo Sample Application was deployed.

kubectl get pods

The BookInfo sample application deployed is composed of four microservices:

• The Productpage microservice is the homepage, populated using the Details and Reviews microservices.
• The Details microservice contains the book information.
• The Reviews microservice contains the book reviews. It uses the ratings microservice for the star rating.
• The Ratings microservice contains the book rating for a book review.

The deployment included three versions of the reviews microservice to showcase different behaviour and routing:

• Version v1 doesn’t call the ratings service.
• Version v2 calls the ratings service and displays each rating as 1 to 5 black stars.
• Version v3 calls the ratings service and displays each rating as 1 to 5 red stars.

You can access the deployed demo application at https://[[HOST_SUBDOMAIN]]-80-[[KATACODA_HOST]].environments.katacoda.com/productpage

In this case, only two of the Microservices, product page and details, are deployed, meaning no ratings will be returned. We'll test the changes to introduce the rating service in a separate namespace and use the Canary release functionality to slow move the traffic.

In the current BookInfo deployment the application is only using V1 of the Reviews service. This means it's returning no rating information.

Before releasing V2 and V3 of the book info, it's important to ensure that it is working as expected. As such we'll deploy the application into a separate namespace, before shifting the for production.

kubectl create namespace bookinfo-pr-7

kubectl label namespace bookinfo-pr-7 istio-injection=enabled

Verify Istio sidecar injection is enabled for bookinfo-pr-7.

kubectl get namespace -L istio-injection

The fully updated version can be deployed with:

kubectl apply -n bookinfo-pr-7 -f https://katacoda.com/aspenmesh/scenarios/deploying-experiments-canary-releases/assets/bookinfo.yml

You can see the new version of the application and running pods using:

kubectl get pods -n bookinfo-pr-7

To control traffic flow within the system, Aspen Mesh supports Developer Experiments.

The experiments allow you to access the new deployment via cookies, or a percentage of traffic.

Task

The experiments are managed the CLI. The experiment defines the target service, in this case the default productpage service and the target service that will receive the traffic, in this case the productpage service in the bookinfo-pr-7 namespace. The TOKEN variable has been pre-populated for you.

aspenctl --token $TOKEN ensure experiment --name productpage-experiment-pr7 --set-service "default/productpage:bookinfo-pr-7/productpage"

By default, nothing changes when an experiment has been deployed. Instead, additional capabilities have been made available.

Within your Aspen Mesh dashboard a new experiment will appear called bookinfo-pr-7 which is what we just deployed.

Within the experiment, a cookie ID has been assigned, similar to the TOKEN needs to be set as a

echo "Enter the Cookie ID [ENTER]: "; read COOKIE

Once the cookie has been set, you can use cURL to access the website. Without the cookie, you should see a response without any ratings mentioned.

curl -s https://[[HOST_SUBDOMAIN]]-80-[[KATACODA_HOST]].environments.katacoda.com/productpage |grep -a5 "full stars:"

When the cookie is passed then the rating service should be called and accessed. You will see the stars being returned in the HTML response.

curl -s -H "Cookie: $COOKIE" https://[[HOST_SUBDOMAIN]]-80-[[KATACODA_HOST]].environments.katacoda.com/productpage |grep -a5 "full stars:"

It's possible to do this via the browser. Access the Product Page at https://[[HOST_SUBDOMAIN]]-80-[[KATACODA_HOST]].environments.katacoda.com/productpage

Using the developer console, set a cookie with your token using the following syntax:

document.cookie='dev-experiment=MYTOKEN; Path=/';

To build additional confident, production traffic can be shifted to the new deployment to see how it behaves.

Within the dashboard, it's possible to send a percentage of production traffic to the Canary Release by adjusting the slider.

Task

Visit the dashboard, adjust the slider and save the changes.

Now, you should see the changes on the product page with the rating service being used.

https://[[HOST_SUBDOMAIN]]-80-[[KATACODA_HOST]].environments.katacoda.com/productpage