Michael Mentele

learning-review   worth-reading

Book Review: The Phoenix Project

Book Author: Gene Kim, Kevin Behr

Summary This is a fun fictional story that is used as a teaching device. It shows how ideas from lean manufacturing can be applied to devops and development.

I especially enjoyed the three ways, here’s a nice recap https://blog.sonatype.com/principle-based-devops-frameworks-three-ways:

  • principle of flows
    • keeping work visible aka kanban
    • limiting WIP so resources can be available
    • reducing batch sizes
    • eliminate, automate, or reduce handoffs
    • identify and eliminate waste and constraints
  • principle of feedback
    • fail fast – tight feedback
    • pushing quality close to the source
    • optimize for downstream work
  • principle of continuous learning
    • it’s okay to learn and make mistakes
    • institutionalize improvement – part of everyone’s daily thinking
    • bubble up improvements
    • test the system – introduce failures, rehearse downtime
    • enforce a learning culture

This was an amazing book to give language to principles I’d been applying and thinking about. The principles of flow I was already bought into and had been exposed through through CI/CD, devOps, and constraint theory. The principles of feedback I was also familiar with – the tight feedback is more obvious. Pushing quality to the source and optimizing for downstream work centers were principles I’ve long advocated for. This is why testing cannot be outsourced to another department such as QA. First and foremost it must live with the developers.

Optimizing for downstream work means working with the business, making sure that when we discover needs we involve engineering and design so they understand the problem. There is so much, so so much work I’ve eliminated in design by translating existing functionality to support new needs. Often we don’t need a brand new widget for example, an existing widget will do. Many times the design is blind to the engineering costs. One design vs. another may be 5% better but 50% more effort to implement. It’s vital we get that handoff right. That is one of the costs of design vs. having engineers do it.

What a fun book. At LeftLane we institutionalize learning via forums, a monthly book club, and regular engineering forums.