If you're reading this, you probably like Kubernetes. Or not. I don't know; I'm just a README file. This repository contains starter code for two servers. In a nutshell, your goal is to deploy them and make them talk to each other.
The first server is in the entrypoint folder.
You can think of it as a user-facing API server.
It should be on Kubernetes and available to the public Internet.
It has the second server as a dependency.
The second server is in the generator folder.
You can think of it as a backend server with high compute requirements.
Since cloud costs (in this hypothetical scenario) can get expensive, we DO NOT
want this on Kubernetes.
There are a couple of dependencies that you'll need:
pythongodockerkubectl
This takehome assumes working knowledge over all of these.
There are many ways to solve this takehome. The actual code and implementation matters less than the end behavior. Here are the specific conditions we require:
- The
entrypointserver must be deployed to Kubernetes. - The
entrypointserver must be publicly reachable. - The
generatorserver must NOT be deployed to Kubernetes. - The
generatorserver must NOT be publicly reachable. - The
entrypointserver must be able to reach thegeneratorserver.
You will be expected to demo this and show that it works. (Also, please don't be cheeky and put everything on the same machine.)
- A public demo URL.
- Any code changes you made to get this working. (A patch file works great.)
- A brief write-up on your approach. Make sure your writing's clear and concise. It doesn't have to be long; just get to the point.
- When developing with Kubernetes, you can use
minikubeor some local single-node cluster. At least that's what I did when creating this takehome challenge. - Feel free to modify any of the code as necessary. The solution is not meant to be prescriptive in any way.
- Reach out if you have questions. This is meant to be a collaborative challenge, not some form of medieval torture.