Contributing a Terraform Guide for Zero-to-JupyterHub-k8s

Hi all, I’m Sebastian Alvis. I work on the Pangeo Project at UW and recently put out the blog post Deploying JupyterHub-Ready Infrastructure with Terraform on AWS. I’m looking to contribute a guide on how to set up a minimal EKS cluster for Zero-to-JupyterHub-k8s, but would like some guidance on what this minimal guide would look like.

In the blog post’s attached repo, I supplied the Terraform configuration files necessary to set up the infrastructure, but this infrastructure had a lot of opinionated design choices that were several steps up from the bare minimum. I think @choldgraf has maintained that Zero-to-JupyterHub-k8s should be minimal but include links to further resources. I plan to make a smaller set of resources for this guide. However, since my blog post used a set of Terraform modules (terraform-aws-eks and terraform-aws-vpc) to configure many of the pieces without a lot of input from me, it may deploy more than the minimal set of resources needed. Is that a fine place to start, or do we need a minimum deployment resource set?

As I work through creating the deployment and guide, I’ll use this thread as a place to drop more questions and I would appreciate feedback. I also plan to make a guide for deploying the JupyterHub via Terraform, but that is a bit further into the future.

4 Likes

This looks great! It’s definitely valuable to have real examples of production JupyterHub deployments in addition to the minimal walkthroughs.

One thing I noticed is that the blog post is a premium Medium post, this may limit who can read it.

There is https://zero-to-jupyterhub.readthedocs.io/en/latest/community/index.html to which I think you could already add a link to the existing repository. That is a mechanism for z2jh to link to other resources that are out there.

Maybe the second repo (the more minimal one?) could even be a second link in that section.