Keywords : AWS, Lambda, EKS, k8s, Lambda Layers, Secrets Manager

Authors : Arunava Kar, Tejaswini B, Arpit Malani (Platform Team @Amagi)

In this post, we’ll attempt to learn kubernetes jobs by actually deploying one. We’ll deploy a simple hello world job using a lambda function with credentials having limited access to the kubernetes cluster.

First, a quick walkthrough of the natively available solutions. AWS’s Batch and Fargate are some of the trivial solutions when it comes to long-running async jobs. These are absolutely great and serve most of the business purposes until we want the application to be on multi-cloud or even on-premise. Let’s see how we can solve this using k8s jobs. The application here can work with EKS, GKE, AKS, on-prem clusters.

This post assumes we already have a running k8s cluster, let it be on any cloud, and have basic knowledge on k8s resources like RBAC, jobs

So, We’ll be deploying the following

  • RBAC role and service account with access to create jobs on k8s cluster
  • Python lambda function to create jobs

#k8s #lambda #aws #serverless

Best Practices for Deploying Kubernetes Workloads in the Serverless Era
1.40 GEEK