No Best Practice In DevOps

I was reading this interview the other day with PagerDuty’s DevOps Advocate (what a job title, eh?) Julie Gunderson which is utterly spot on in how it addresses culture and DevOps all the way to how it mentions the three capital pieces of reading I keep obsessively sending people to these days - the Phoenix Project, The DevOps Handbook and my beloved recent DORA State of DevOps report. 

While the interview is excellent and needs bookmarking, one thing I found irritating is the fact that it mentions “common challenges and best practices”. 

On a second read it’s evident the terminology is used by the publication not Julie, as she elegantly diverts and reframes and at times bluntly answers with “invest in platform and people, that’s what is best practice” but the reality is that the term is there and that it’s firmly etched into not only our mindless business everyday vocabulary but our mentality and our definition of success. 

I’m not the only one who has been irked by the term “best practice” for a while. I heard many others complain be it from the stage, in writing or in conversations. It is an antiquated, inapplicable term that has no place in today’s work environment that moves way too fast and in far too personal of ways for there to be any general sweeping generalisation of what works.

Let’s talk about why a culture of asking and searching for “best practices” when it comes to #Agile mindsets and DevOps success is bankrupt. Before we do, let’s also agree when the demand comes, it isn’t about the technology in itself, it does refer to culture and ways of doing things and it is never on the exact architecture and software used in the creation of the CI/CD pipes. 

It’s chiefly because there is no best practice in VUCA. 

There is no best practice in DevOps. 

There is no best practice people wise. 

There are fundamentals in terms of technology and mindset and there is a whole lot of learning to be done. There’s the Agile Manifesto and there’s common sense, empathy and goodwill. There are some amazing stories of companies that have done some kick-ass things and there are some awesome Superheroes who have made those happen.

But no best practice. 

The “elite performers” in the report are no models. They can not be copied. They haven’t themselves xerox-ed Google’s organization or practices and made it. There is no DevOpsInsta where you just add water and you’re on top. Everyone who is doing well put in extremely hard work and bespoke amounts of thought. All best practices were harmed in the making of their success. 

Let’s be honest - anyone who asks for them is:

• Stuck in the past. With no real understanding of the fact that things are simply not having the time to become “best practice” - methods, tries, MVPs, none of them perennial or fixed or applicable enough to become examples.
• Lazy. Let’s face it, asking for exact models to copy is mainly powered by sheer unwillingness to do the work of creating the best culture and platform that are needed.
• Afraid. It’s mainly risk aversion - this need for silver bullets. And it’s not strange, most companies have bred a culture of protecting each milestone and recognisable successful action but have done so in a time where stasis was possible, no longer the case. Courage is non-negotiable these days.

This doesn’t make the ones who ask for best practices intrinsically bad and I strongly believe most businesses are comprised of humans who have immense reserves of passion and curiosity buried deep under those demands, though so we can and ought to help them access them and stop being the above if we change the narrative and overtly and categorically refuse to entertain it. 

So if they can’t have “best practices” what can they have? 

• Learnings. Lessons, not models. Looking at others and their journey with DevOps and applying what may work while contemplating what wouldn’t and being willing to then re-examine that lesson and learn again and again.
• Food for thought. While there are no prescriptions, everything elite performers do is interesting to consider and opens new dialogue avenues and every topic should be up for discussion and dissected to build this new high performing business reality.
• A relentless focus on people and interactions. Not an original one and no news that the lens should not be on studying the process but on better understanding one’s own people and interactions but the fact that the focus in itself can replace the need for best practices, is what’s new. The preoccupation alone with what it is that makes your own culture tick by having your own teams psychologically safe (more on how culture magic is but a set of teams that work well in an article to come) is more likely to pay dividends than any examples or models. 

I think it’s time we all do admit that we have to take a harder stance against it. The mere demand from the business is something we all perpetuate by allowing the ask. 

In fact the DORA report acknowledges this directly “Many organisations wanting to adopt DevOps look for a set of prescriptive steps or best practices to guide their journey” but stops short of tackling it to say “Enough is enough, when the business asks for it, just say “There is none, get over it, don’t be stuck, lazy and afraid, we’ll be our own best practice and then we’ll be a better one after that”.