A single dramatic software failure can cost a company millions of dollars — but can be avoided with simple changes to design and architecture. This new edition of the best-selling industry standard shows you how to create systems that run longer, with fewer failures, and recover better when bad things happen. New coverage includes DevOps, microservices, and cloud-native architecture. Stability antipatterns have grown to include systemic problems in large-scale systems. This is a must-have pragmatic guide to engineering for production systems. — via PragProg.
Michael Nygaard wrote the first edition in 2007. I read it a fews later and it drastically improved my approach to software engineering. The book was especially helpful at that point in time because I didn’t have real experience operating production systems. It was like a preview for what lay ahead and it prepared me ore it. Much as changed for me and the IT landscape over the last decade. Release It! (Second Edition) ports the lessons and perspectives from the 1st edition to the modern landscape of cloud computing, DevOps, continuous deployment, containers, microservices, and more.
Software deliver its value in production. The development project, testing, integration, and planning…everything before production is a prelude. This book deals with life in production, from the initial release through ongoing growth and evolution of the system.
This quote from the introduction eloquently distills the approach better than I can. Release It! is all about building software that succeeds production. Production is a dangerous place. These days any web service can see millions of users. Like it or not, Facebook has a billion users. Architecting systems and designing code to function in these conditions is a challenge.
The book is a four part roadmap for improving production operations:
- Create Stability
- Design for Production
- Deliver Your System
- Solve Systemic Problems
Readers familiar with the first edition will recognize the case studies. The airline grounded by an uncaught SQLException is painful and still relevant.
I am happy to report that that 2nd edition moves away from a Java heavy narrative to one one more grounded in cloud infrastructure, open source, and DevOps practices. It’s no surprise that 1st edition took this route. Things are vastly different now. There are three major cloud providers now, DevOps is main stream, there are a plethora of more languages in production (Node.js comes to mind), and containers have taken infrastructure by storm. The second edition prepares you to build production ready software with modern tools.
Congratulations to Michael Nygaard on his writing! Nygaard comes across as someone who has been through it all — from writing code in the trenches, fighting outages, and architecting systems to survive it all. He’s also been on the receiving end of outages that cost real money — a fact that’s not lost throughout the text. His writing is engaging, comical, and honest. This is refreshing in a technical book.
Everything was on real hardware, no virtualization. Just melted sand, spinning rust, and the operating systems.
Nygaard always seemed to elicit a chuckle or jokingly date some examples in the book. His telling of outage war rooms stories create bonds with readers in similar situations. I bet that readers will relate to carrying a laptop to family gatherings or going dark at exactly the wrong time. His advice doesn’t come from an ivory tower. It comes from first hand experience.
Release It! recommends a philosophy for building production ready software. Failure is inevitable, so it must be accounted for and planned for. The goal is not to eliminate failure, but promote durability and quick recovery times.
The book feels like an extension to the DevOps Handbook. The DevOps Handbook provides three ways for building high performing IT teams:
- The Principle of Flow (fast flow from development to production)
- The Principle of Feedback (fast flow of production operational state to development)
- The Principle of Continuous Improvement (continuing improving the previous two feedback loops by building organization and culture habits).
Release It! touches on all three ways. Part one (“Building Stability”) relates to the principle of continuous improvement. Businesses cannot move forward if production is constantly on fire. Part two (“Design for Production”) relates to principle of feedback, or building telemetry into production systems so informed decisions are possible. Part three (“Deliver Your System”) relates to the principle of flow in designing systems for continuous delivery/deployment. This is the crux of principle of flow. Part 4 (“Solve Systemic Problems)” connects with the principle of continuous improvements. Release It! advocates chaos engineering as a means (and mindset) to routinely remove failure points and increase reliability.
It’s no surprise that DevOps principles and practices surface in Release It! I don’t think the author explicitly states that DevOps is the best way to build, ship, and run software. However it certainly feels like there’s a strong correlation if you’re familiar with the topic.
The chapter on Chaos Engineering is welcome addition. I couldn’t fathom this in the 1st edition. It’s another indicator how much the IT landscape has changed and how we oriented ourselves in it.
Release It! contains a wealth of valuable information for teams and individuals. This point sticks with me. Pay close attention to integration points with any system. They will fail and take you down with them if you’re not careful. Oh, and watch your thread pools and queues!
Highlights and Take-aways
Every failing system starts with queue backing up somewhere
This quote refers to TCP listen queues. It’s simply good insight. There are queues (likely they’re implementation details to your abstraction layer) in front of everything. Things go hang wire when they fill or overflow. ELB surge queue anyone?
Design and architecture decisions are also financial decisions. These choices must be made with eye toward their implementation cost as well their downstream costs. The fusion of technical and financial viewpoints is one of the most important recurring themes in this book.
I appreciate that Nygaard drives this point throughout the book. Outages cost real money. An outage destroyed Knight Capital. An outage may cost you $1,000,000 one day. The synergy between these factors cannot be ignored.
The main lesson is that not every problem can be solved at the level of abstraction where it manifests. Sometimes the causes reverberate up and down the layers. You need to know how to drill through at least two layers of abstraction to find the “reality” at the level in order to understand problems.
This quote comes from a scenario debugging application problems originating in the networking layer. Systems sit on increasingly higher abstraction levels, but they’re still exposed to failures in lower levels. This is a good reason for reading and understanding code at lower levels. In other words: if you use a framework such as React or Rails, read the code and gain understanding.
A new architect will focus on the boxes; an experienced one is more interested in the arrows
This quote is in parenthesis in the text, but I found it more valuable than the sentence itself! Release It! is full of these quips. They show me Nygaard depth of knowledge and experience. More importantly, they point less experienced readers in the right direction.
Hope is not a design method
This felt like a direct reference to the SRE book’s “Hope is not a strategy” mantra. This line of thinking makes Release It! a valuable read for those in the SRE space.
In time, even shockingly unlikely combinations of circumstances will eventually occur. If you ever catch yourself saying “The odds of that happening are astronomical,” or some similar utterance, consider this: a single small service might do ten million requests per day over three years, for a total of 10,950,000,000 changes for something to wrong. That’s more than ten billion opportunities for bad things to happen. Astronomical observations indicate there are four hundred billion stars in the Milky Way galaxy. Astronomers consider a number “close enough” if it’s within a factor fo 10. Astronomically unlikely coincidences happen all the time.
Just putting a pin in that one then.
Transparency arises from deliberate design and architecture. “Adding Transparency” late in development is about as effective as “adding quality”. Maybe it can be done, but only with greater effort and cost than if it’s been built in from the beginning.
Transparency here refers to production transparency provided by telemetry systems. This is another common principle in DevOps workflows: build production operational concerns in as soon as possible. Things improve from there.
The only real answer here is do your homework and commit to solving implementation challenges with whatever tool you choose.
This quote comes from a discussion on choosing a monitoring system. It applies any big technology choice. There’s no silver bullet. At the end of the day, you must know what you’re getting yourself into, then learn to use the tools to get what you need done. Nothing will fulfill 100% of your needs out of the box, and even less likely, as things change over time. Just choose, commit, and run with it.
Release It! (Second Edition) is required reading for software engineers looking to improve their next or current production system. It’s especially useful to those wearing many hats. Personally, I think less experienced engineers will gain more than experienced engineers. This book is definitely for technical minded folks, so don’t bother recommending it to your CEO or other upper management if you’re not already having these conversations with them.