Description
The DevOps methodology is reality. That said, many developers have to deal with the process of bringing python applications into production. One aphorism of the Zen of Python states: "There should be one-- and preferably only one --obvious way to do it."
While for the language itself and code being written with it, this advice is often strictly followed for good reasons. For the process of bringing python into production it is not.
In the first part of this talk I will give an overview of the different steps in the delivery pipeline, from packaging to deployment and the various possible implementations for each of those steps that emerged over the last years. We will discuss docker, pex, wheels, debs, and tars. We will learn why fixed environments are crucial and why pip is not yet what it should be. We learn about the different runtime environments, from bare metal servers over ephemeral container clusters to “serverless” and what they impose on the deployment process and scratch on the surface of configuration management using tools like Ansible and orchestration frameworks like Kubernetes. We will also explore the world beyond python by looking into other languages ecosystems and learn how they addressed and solved this issue.
After we understood the requirements and realized that there is not "one obvious way to do it", this talk is a call to action: Let us define and build the “one obvious way” of how we want to bring python into production!