Re-Thinking Workflows : Development and Maintenance of Drupal Instances at Scale

For the past several years, we have been developing and managing a moderate amount of Drupal instances across several VM hosts using a homegrown product we call pushback. It leverages drush in conjunction with Jenkins and gitHub. Why develop a custom framework? Although several products existed [1] [2], none seemed to meet our desire for simplicity and maintainability.

Pushback was born through discussions with Jen Whitney and Kassim Machioudi, and was inspired by the (now-abandoned) Tugboat project by Lullabot. It is simple and does nothing novel. A summary:

This workflow worked very well for our small team. Some positive benefits:

It worked great. It did what we needed. But, as we began to challenge our workflow and infrastructure in the past 18 months, several general discussions kept recurring:

These are difficult questions, and introduce problems not easily solved. I feel the recent rise in popularity of a container-centered infrastructure and workflow offers strong possibilities for the future.

Post Tags:


comments powered by Disqus