Skip to main content
How is your Kubernetes (K8s) setup configured on AWS? Please explain. This question is often used at the beginning of interviews to understand the specifics of your Kubernetes environment. Interviewers are interested in learning whether your Kubernetes cluster is deployed directly on EC2 instances, managed through a service like AWS EKS, or provisioned using other tools.

​
Approaches to Setting Up Kubernetes

There are several common methods to deploy a Kubernetes cluster:
  • Local Development:
    For development and learning purposes, many opt to use Minikube for its simplicity.
  • Cloud Environments:
    When deploying in the cloud, popular approaches include:
    • Managed Service:
      AWS EKS (Elastic Kubernetes Service) is a managed solution optimized for production workloads, reducing maintenance overhead.
    • Self-Managed Clusters:
      Utilizing EC2 instances combined with provisioning tools like Kops or Kubeadm to set up a Kubernetes cluster.
For organizations using other cloud providers, similar managed services exist:

​
Production vs. Development Environments

In production environments, the focus is on stability and efficient maintenance. Managed services, like AWS EKS, provide a significant advantage in scalability and integration with other AWS components. In contrast, self-managed clusters using tools such as Kops or Kubeadm generally involve more operational overhead and are more suitable for staging or development purposes. In our current setup:
  • Production:
    We operate a managed AWS EKS cluster. This approach minimizes maintenance and ensures high compatibility with AWS services.
  • Staging/Testing:
    For non-production environments, we use Kops to rapidly spin up Kubernetes clusters, which offers a quick and flexible setup.

​
Example Workflow for Deployments with Helm

Below is an example command that demonstrates how to deploy an application using Helm: 
# Example deployment command using Helm:
helm install my-application ./my-chart
If you have experience with GCP or Azure, simply substitute the AWS-specific services with their counterparts (GKE for GCP and AKS for Azure).
By outlining your Kubernetes setup in this structured manner, you not only demonstrate a robust understanding of both production and development environments but also emphasize the operational considerations essential for modern cloud-native applications. That concludes the discussion for this article. Thank you, and see you in the next lesson.

Watch Video