Laboratorio Multi-Cloud

How to manage a multicloud infrastructure?

Working in a multicloud environment can be daunting. Not only does it mean using multiple cloud services, but also ensuring their interoperability. In this lab, we will learn how to manage infrastructures using Google Cloud Platform (GCP) and Amazon Web Services (AWS). Let's see how to start provisioning, configuring and visualizing our applications in a multicloud environment.

What do we need to get started?

To conduct the lab, you need to have:

• Credentials to access GCP and AWS.
• A GitLab repository with the Anthos Multicloud workshop.
• The Quick Labs platform, which provides us with credentials and some preconfigured infrastructure.

How to configure working environments?

We start by logging into the GCP console(console.cloud.google.com) using the provided credentials. It is important to choose the right project using gcloud config set project. This ensures that we work in the right environment.

gcloud config set project <YOUR_PROJECT_ID>

Next, we configure the environment variables that will be needed to provision the infrastructure on AWS. These include the secret_access_key, AWS_access_key_id, among others. Next, we create a directory for our project with mkdir and prepare it to work there:

mkdir ~/AnthosMulticloudcd ~/AnthosMulticloudexport  WORKDIR=~/AnthosMulticloud .

How to provision infrastructure?

The provisioning process uses scripts that have defined instructions to deploy the necessary infrastructure. We use Terraform to provision resources in GCP and AWS in different environments: development, staging and production.

git clone <REPO_URL>cd AnthosMulticloudWorkshop./build.sh

We monitor the process from Cloud Build in the Google console, where we will see how several jobs are deployed to lift the infrastructure.

What services are deployed?

The deployment creates:

• GKE clusters for Google Cloud.
• EKS clusters on AWS.
• A GitLab server for remote configurations.

During the provisioning script, tasks such as cluster creation and networking configuration are performed for each environment. This process can take 20 to 25 minutes.

How to visualize the deployment?

To ensure that our infrastructure is working properly, we access the deployed services and review their performance. We use kubectl to verify the Kubernetes clusters, and Kiali to monitor the traffic and status of the services in real time.

kubectl get pods -n istio-system

Kiali provides a visual graph showing how services connect to each other and the flow of traffic between microservices.

How to handle service portability?

One of the essential aspects of the multicloud approach is to move services between clouds quickly:

1. If a service fails in AWS, we can migrate it to GCP.
2. The deployment script helps us create a new service in GCP.

For example, if the cart service in AWS stops working, we recreate it in GKE, which restores the application without the end user noticing any interruptions.

How to deploy multiple instances?

Finally, we deploy different instances of the front end in both GCP and AWS clouds using tags and selectors. This allows load balancing between clouds and ensures high availability. When accessing our application, we can switch between instances on GCP and AWS without interrupting the user experience.

With these configurations and scripts, we have achieved a robust multi-cloud environment that is ready to handle traffic and contingencies effectively. Continue to explore and gain hands-on experience to become proficient in multi-cloud environments - you have everything you need to continue learning and improving your skills!