New Year’s is a time to reflect on past successes and how to improve in the coming year. Such retrospection is probably not the first thing that comes to mind when thinking about engineers who, according to movies, ought to be staring at log lines whizzing down the screen at considerable pace.
Over the course of last few years, Mapzen engineers have amassed an amazing stack of systems and software that support our public facing projects and APIs. We will be publishing a series of blog posts about our engineering practices, including what has succeeded and where we have failed. Here we’ll cover various application frameworks. We’ll go over AWS and the services we utilize in our attempt to reach what we like to refer to as ‘AWS bingo’. We’ll describe the way we manage infrastructure complexity and setup with Chef or Docker as well as some of the tooling we have built around it. We’ll also discuss API management as well as metrics both for business analytics along with traditional elk stack deployment. We’ll even talk about how we handle system outages and problems, although everything we do works perfectly all the time. ;)
Mapping is quite well covered on this blog, but this series is about the not actually mapping engineering (NAME). Our motivation is to discuss topics that are not necessarily related to Tangram, search, navigation or map tiles, but rather the cradle that supports them.
We deeply believe in open source and keep as much of our code open and available as we can. We’ve tried to maintain a standard where source code that addresses common problems is shared. Some things we do keep private, such as Mapzen-specific code or things that are undocumented, but in this series we will talk about them too.
We hope that by sharing our methods we can cut short someone’s troubleshootings efforts when implementing large scale systems or building a mobile SDK. We would not be here without open source, and we want to be able to give back. You will be able to jump directly to this series via /tag/engineering.
The first post by Grant Heffernan will be on building a data pipeline to Elastic Search to keep Mapzen Search data fresh and available.
