The DevOps Handbook: How to Create World-Class Agility, Reliability, & Security in Technology Organizations is a practical roadmap for building high performing IT organizations. It’s written by well known members of the DevOps community: Gene Kim, Jez Humble, John Willis, and Patrick Debois. They distill practical information and experiences in moving towards a DevOps value stream. This review includes the background information on the book, my thoughts, favorite passages, and a recommendation. If you’re solely interested in the recommendation than skip to the end.
Here’s the book’s pitch:
Increase profitability, elevate work culture, and exceed productivity goals through DevOps practices. More than ever, the effective management of technology is critical for business competitiveness. This non-fiction follow-up to The Phoenix Project shows leaders how to replicate these incredible outcomes, by demonstrating how to integrate Product Management, Development, QA, IT Operations, and Information Security to elevate your company and win in the marketplace.
The Phoenix Project is a fictional story about a struggling company and their transformative success. The DevOps Handbook is manual for putting the changes in The Phoenix Project into practice. The author’s make strong claims about increasing probability and success in the marketplace. Naturally all business desire those outcomes. The books is light on the data side, but rich with personal experiences, case studies, and anecdotes.
The author’s split DevOp’s practices into three ways:
- The Principle of Flow: fast flow from left (idea/code) to production (right)
- The Principle of Feedback: fast information flow right (production) to left (code)
- The Principle of Continuous Improvement: building organizational feedback loops to keep things moving
DevOps encompasses far more than simply breaking down the barrier between development and operations. It’s a philosophy built on feedback loops. Each of the three ways is a feedback loop in its own way. First, lower lead times with continuous delivery so ideas make it to production faster. Second; improve telemetry throughout the process so failures are more quickly recognized and resolved. Third: build organizations that leverage these feedback loops and empower everyone to improve in their daily work. Four: (yes) profit.
The book is divided into three parts, one per way. Each part documents the relevant technical practices with supporting case studies and experiences sourced from industry professionals. It’s no surprise that companies like Netflix, Amazon, and Etsy are referenced often.
The author’s note that three ways are not a linear progression. Every team may have bits and pieces of each. This is where the DevOps handbook shines. It draws a roadmap across them all so improvement focused teams may fill their gaps.
The DevOps Handbook is fantastic. I poured across the pages and found many noteworthy passages. There is a ton of valuable knowledge and insight in the book. Reading it put my feelings around DevOps into words backed by statistics and practical implementation patterns.
It’s no secret that I think that continuous deployment is the best way to ship software. The author’s focus on continuous delivery, but there’s enough overlap between them and the other practices they bring along for the ride. I couldn’t help but nod along through part one until the continuous integration chapter.
Continuous integration, like many terms in software development, means many different things to many different people. CI, to me, means running automated tests against all commits to source control. My definition is orthogonal to anything besides automating testing. My definition of CI refers to practice of automated testing in the DevOps Handbook. The DevOps Handbook defines continuous integration as the combination of trunk-based development and automated testing.
Roughly speaking, trunk-based development requires developers to check in code multiple times a day to trunk (or “mainline”, or “master”). This gives me pause. My immediate thought was there’s no way to do away with topic branches. I remember the branching strategy rules in my past teams and those published in blog posts. You may be familiar with “git flow” with it’s release, hotfix, and jumble of other purpose built branches. This approach trips many people up because branching strategy is interwoven with how they physically do their job. It seems this a common reaction because the author's devote a small section to addressing these concerns immediately after introducing the concept.
They acknowledge trunk-based development is the most controversial practice they advocate for. I imagine the controversy stems from the fact that trunk-based development drastically changes how people do their work. TDD, and CD faced the same resistance when they were introduced. CI may be the last domino. The authors do their best to dispel the controversy with experience and case studies. Their position is that branches are useful. Long running and complex branches are not because they have a negative impact on productivity. Trunk-based development mitigates the riskby reducing batch sizes (since smaller branches are easier to write, test, and integrate).
The CI chapter made me revaluate some past decisions and their local and wider organizational effects. I’ll take trunk-based development more seriously from this point on.
Two terms from the DevOps Handbook resonate with me: lead time and batch size. It’s no surprise that “lead time” and “batch size” are peppered across this review. Lead time is the time to go from customer request to code running in production. Batch size is roughly equivalent to the size of a particular change. Of course the two are connected. Smaller batches (such as 20 line method change) are easier to develop, test, and deploy. This reduces lead times building an important feedback loop. You’ll find these terms mentioned throughout the text. Both are helpful in framing your (or anyone else’s) organization’s problems.
The problems facing you (or your organization) are likely technical and organizational. The authors hammer home the point that DevOps is not simply technical practices. It’s the sum of organizational practices and technical implementations. There’s plenty of focus on the organizational side throughout the text. I didn’t expect to read about Target here, let alone be impressed by their DevOps dojo. That just goes to show that any organization can do DevOps and do it well.
Highlights & Take-aways
Admittedly I have a hundred or so notes and highlights. Here are the ones that resonated with me the most.
In DevOps, we typically define our technology value stream as the process required to convert a business hypothesis into a technology-enabled service that delivers value to the customer.
This passage made me change how I wrote and talked about DevOps. “Business hypothesis” sounds a lot like lean. Lean and DevOps dovetail into each other nicely.
In addition to lead times and process times, the third key metric in the technology value stream is percent complete and accurate (% C/ A). This metric reflects the quality of the output of each step in our value stream.
I had not encountered this metric until the DevOps Handbook. I realized that precent complete/accurate would have helped quantify problems in my previous organization.
Internally, we described our goal as creating “buoys, not boundaries.” Instead of drawing hard boundaries that everyone has to stay within, we put buoys that indicate deep areas of the channel where you’re safe and supported. You can go past the buoys as long as you follow the organizational principles. After all, how are we ever going to see the next innovation that helps us win if we’re not exploring and testing at the edges?
This is my favorite passage from the book. I used to be a lifeguard so the analogy really resonated with me. Teams need structure to succeed in common work and the freedom to break the mold when the situation calls for it. More importantly you have to trust the team to wade past the buoys and handle the risks accordingly.
The implications of the Kohavi data are staggering. If we are not performing user research, the odds are that two-thirds of the features we are building deliver zero or negative value to our organization, even as they make our codebase ever more complex, thus increasing our maintenance costs over time and making our software more difficult to change. Furthermore, the effort to build these features is often made at the expense of delivering features that would deliver value (i.e., opportunity cost).
There’s the stark truth for many product organizations where plan according to gut feel. My take-away: don’t guess; know. Repeated poor guesses reduce moral and eventually lead to burnout — or worse.
As Mike Rother wrote in Toyota Kata, “As tempting as it seems, one cannot reorganize your way to continuous improvement and adaptiveness. What is decisive is not the form of the organization, but how people act and react. The roots of Toyota’s success lie not in its organizational structures, but in developing capability and habits in its people. It surprises many people, in fact, to find that Toyota is largely organized in a traditional, functional-department style.”
This passage hammers the point that people’s habits and practices are far more important to long term success than organizational structure.
I don’t have a bad word to say about the DevOps Handbook. It’s a practical roadmap to improving IT in any organization. It’s also the most valuable book on software development I’ve read in the past 10 years. I think anyone curious about DevOps or improving their organization should read this book. I’d be shocked if you didn’t find anything valuable.