PuppetConf 2017

Did you know there are more than 500,000 Puppet files in open source repositories on GitHub? That's more than 7 million lines of Puppet code. What if we were to analyse all of that code? What could we find out about how people write Puppet in the real world? And what would those findings mean for anyone using or developing Puppet today? In this talk we'll demonstrate how to go about analysing this corpus of data using Google's BigQuery service and other tools. We'll ask questions like: * What are the most, and least, popular resource types? * What are the most, and least, popular built-in Puppet functions? * How many people are using the Puppet 4 data types? * What are the most common module dependencies? * How common is the hiera function vs relying on data binding? The purpose of this talk is to shine a light on popular patterns and practices within the Puppet community. We'll also show people how to conduct the same analysis on their own internal and private Puppet repositories. From that understanding we can build better tools, or argue more convincingly for features or better identify areas of improvement, both in our own code and in the wider ecosystem.