francoposa.io

/home /resources /about
Home > Resources > Infrastructure & Operations

Zero to Production with Kubernetes, Part 0: Goals and Tool Selection

Software Infrastructure and Operations with DigitalOcean Droplets, Ansible, K3s, and Helm

Franco Posa

Purpose

As software engineers, the early months and years of our careers are often highly focused just on writing working code. Then, as soon as we have our heads above water, we immediately seek (or are asked to) to write clearer code, more impactful code, more performant code, and to solve increasingly complex technical or end-user problems.

All of this coding and technical education and growth is often confined within the application boundary - a mobile app, a web frontend, a backend CRUD server, or a data pipeline. We rarely learn in any depth - if at all - how to deploy and operate our software in production.

This gap in experience and knowledge limits our ability to take advantage of one of the most unique aspects and greatest joys of software engineering, which is that a single person or a very small team can create, deploy, and operate a nontrivial application or entire system of software.

Let’s get to it.

Goals

This series of guides seeks to a strike a balance between learning and getting shit done. These two options are naturally in tension, and then balance chosen by the guides may not be for everyone. Feel free to skip over explanations when you are in the mood to just get shit done, or to deep dive into the resources and documentation when you are hungry to learn more.

With that said, we will:

1. Ship Software!

We want to give ourselves the ability to take an app from running on our local machine to running on a public cloud server.

Further, we want this deployment process to be simple, automated, repeatable, and not to involve significant customization. That is, the process, tools, and configuration used to deploy one application should be easily transferable to a second, third, or hundredth application.

Specifically, we will deploy a backend HTTP service - a simple echo server - to a Kubernetes cluster running on a cloud VM, reachable over public internet via a domain name.

2. Learn Kubernetes Concepts and Associated Tooling

We are not going to become Kubernetes experts by reading a few guides, nor do we aim to. As stated above, the number one goal is to ship software.

We will:

3. Learn Infrastructure & Operations Concepts and Associated Tooling

Though we will not go super deep on these topics (because we are focused on shipping software), we will gain some basic exposure to:

Tool Selection

Here again, we will try to strike a balance. The tension in software tooling is often between a perceived “simplicity” or “user-friendliness” vs. the tool being “advanced” or “feature-rich”.

The tools below are chosen to allow us to get off the ground easily without needing to peek behind the curtain too much. At the same time, they are powerful enough to allow us to push into much more advanced territory if we desire.

Moreover, software tools are just vessels for overarching concepts and practices which can be applied universally. At this stage of our learning, it is far more important to understand concepts than to become a master of a particular tool. Thousands of software tools will come and go, and at the end of it we will still be managing servers, abstracting the concepts of compute, storage, and networking, and packaging, deploying and monitoring software on those servers.

DigitalOcean

Ansible

K3s

Helm