Getting started with Chef Habitat.
This meetup included a selection of talks, one of which gave me the opportunity to see Habitat in use for the first time.
Habitat is platform designed to support the entire DevOps lifecyce... all the way from building to configuring, deploying and running services. It even allows for some snazzy things like deploying new confogurations to your services at runtime.
For a fuller perspective, here is the original launch anouncement video:
Suitably enthused, I made my way through the "Getting Started with Habitat" tutorial which is easy enough. With some much prepared in advance, I decided I wanted to build and deploy my own service from scratch and so I decided to create a super-simple "Hello, World!" service using Python and Flask.
The aim of my service was simple:
Creating the service is easy enough and then getting Habitat to package it, likewise. There was one small gotcha in that using Python virtual environments inside Habitat caused me many, many headaches (notably, installing dependencies was fine but Python itself would baulk). To be honest, this may very well be PEBCAK. Regardless, I ditched the venv and installed Flask without one.
After this, using Habitat to build and package is super simple.
The final step is then to checkout start-time configuration of the packaged service. Habitat provides great functonality for this (you can see the docs here).
Here you can see how I provide the name "Paul" to a new instance of the service.
So far I am enjoying playing with Habitat. Definitely have a lot more to learn but, even with the little I have covered, it is clear that Habitat is a great tool for putting the entire DevOps lifecycle in the hands of the service developer.
One of the super-sexy features of Habitat is reconfiguration of the service at runtime! That's what I'm looking into next.
In the meantime, please feel free to checkout my work over on GitHub.