I'm trying to deploy Kong, an open-source API gateway and microservices management layer, on Amazon Elastic Kubernetes Service (EKS) using AWS Fargate for container orchestration. However, I'm encountering some challenges with this setup, and I'm looking for guidance on the steps and best practices to make it work.
I've already created an EKS cluster with Fargate profiles, but I'm unsure how to proceed with deploying Kong and ensuring it works correctly.
Could someone provide a step-by-step guide or point me to resources and best practices for deploying Kong on EKS with Fargate, along with any tips to make the setup secure and performant?
Thank you for your assistance!
Specifically, I'm interested in:
- Installing Kong on the EKS cluster.
- Configuring Kong to work effectively within the Fargate setup.
- Any necessary network and security configurations.
For your questions:
Installing Kong on EKS Fargate is not different from installing it on other Kubernetes services.
I'd recommend using Helm to manage your installation, Kong offers an official Helm Chart repository and values to get you started.
I think the main selling point of Fargate services is it's the serverless aspect of it. So deploying Kong with specific request/limit and HPA should be enough and let Fargate do the rest.
This will be entirely dependent on the requirement of your cluster.
Is the cluster private with public NLB/ALB? What's the requirement for DC/DR site? How complicate is your AWS networking?
This blog post should help you see the bigger picture.
Also, using EKS Fargate require some modification on the kube-dns and stuffs.
About using Kong with AWS RDS, if you configured it to allow connection from EKS subnet/vpc and creating user/database for Kong, It'll work just like using self-hosting PostgreSQL.