BT

Facilitating the Spread of Knowledge and Innovation in Professional Software Development

Write for InfoQ

Topics

Choose your language

InfoQ Homepage Podcasts John Willis on DevOps Evolution, Leadership and Burnout

John Willis on DevOps Evolution, Leadership and Burnout

Bookmarks

Key Takeaways

  • DevOps got watered down along the way, but its principles and practices will stay.
  • People can learn the technical side of DevOps with training but they need to follow up on case studies from organizations that went through similar journeys.
  • Still early days to be able to distill what are the good and bad DevOps leadership practices, but we at least know that blameless environments are much more productive.
  • We are missing a burnout survey in DevOps, number of people affected probably staggering high but no one knows for sure, so the problem gets underrated.
  • The new view on human error is that we need to look at the system that allowed people to make the mistakes in the first place.

Notes

How DevOps Started and Its Evolution

  • 1:54 - DevOps started as a codification of good practices from different areas.
  • 2:25 - The first DevOps Days conference in Ghent back in 2009 highlighted a new way of thinking about Ops work.
  • 3:00 - The second one took place 6 months later in Mountain View in the US and gathered 300 people already.
  • 3:25 - Been a DevOps evangelist since then, kickstarting new DevOps Days conferences around the world.
  • 3:59 - Like any movement that becomes popular, it's normal that the term DevOps got watered down and that's ok.
  • 4:25 - Practices and principles is what sticks around after the name loses its original meaning.
  • 4:34 - Digital transformation will be the next hype wave, but we shouldn't throw out DevOps practices.
  • 4:55 - Mainframes had very mature practices around change management which were dismissed when distributed systems became popular. We should not repeat the same mistake.
  • 5:55 - Cross-fertilization between movements is important, just like DevOps comes from Lean and Lean comes from the Toyota system.
  • 6:07 - For example, human factors, safety, resilience engineering are practices and principles that are common to many areas.

Is DevOps Training Needed

  • 8:43 - People need to learn both from case studies and structured training.
  • 9:42 - There's a place for technical training on things like Continuous Delivery or monitoring.
  • 11:40 - Recorded an "Introduction to DevOps" free course for the Linux Foundation.
  • 12:09 - The course includes additional case studies since the DevOps Handbook was published, and also covers base knowledge not in the book.
  • 12:56 - The course goes into a lot more detail about value stream mapping than the DevOps Handbook.
  • 13:06 - The course has multiple versions: core version with 15h of video lectures, advanced version adds 30h of recommended video presentations and case studies, and last version adds recommended whitepapers and books.
  • 13:48 - Recorded a follow up course on introduction to Continuous Delivery with Docker Flow.
  • 15:30 - Toyota's lean model is not well understood in western countries. In perspective it's still quite new.

What Makes a DevOps Leader

  • 15:50 - You learn about leadership by reading the DevOps Handbook, or attending the DevOps Enterprise Summit and talking to people who've led DevOps transformations, or at least watch the videos online.
  • 17:00 - The journeys of leaders like Courtney Kissler (first at Nordstrom, now at Starbucks) are the "real" leadership books.
  • 17:13 - It's still early days to be able to codify what are the good and bad leadership patterns in DevOps, there is no single book or training that can provide that.
  • 18:05 - Leaders come from a variety of backgrounds and thought models, from lean thinking to technically focused.
  • 19:10 - Psychological safety is an extremely important concept that DevOps leaders must understand in order to create an enviroment where failure is embraced.
  • 20:50 - Good leadership is creating a blameless culture.
  • 21:07 - There are different types of cultures: pathological, bureaucratic and generative. The latter treat failures as learning opportunities.
  • 21:40 - Generative cultures tend to be faster and more resilient, according to both survey (statistical) data and empirical data as well.

DevOps Handbook

  • 7:58 - The book talks about Lean ideas such as lead time and lean metrics that are common to many of the case studies coming from the DevOps Enterprise Summit.
  • 8:50 - There are 50 to 100 new case studies from DevOps Enterprise Summit available since the book was published.
  • 10:05 - Working on an audio series with Gene Kim on "Beyond the Phoenix Project" looking at pre-DevOps body of work and origins.
  • 10:23 - The DevOps Handbook is the codification of a model of practices, called the 3 ways: flow, feedback, and continuous learning.
  • 10:49 - Case studies in the DevOps Handbook are mostly from enterprises (30-35) and some (10-12) are web scale.

DevOps Enterprise Summit

  • 9:18 - All the videos/case studies from the conference are available on YouTube, unless the company didn't allow it.
  • 22:25 - Last year's joint talk with Damon Edwards was about "Faster, cheaper, safer DevOps", around how to actually do that in practice.
  • 22:48 - This year's talk is about the production side of flow, how to use immutable delivery infrastructure with Docker, and how to operate it at scale.

Burnout and Human vs System Error

  • 24:15 - An article in 2015 following a young IT professional's suicide due to burnout attracted a lot of attention.
  • 24:45 - In 2016 there were a lot of conversations around burnout but today there's hardly any discussion around this issue anymore.
  • 25:29 - Burnout is a disease, we need to keep awareness about it.
  • 26:10 - Wants to run a "Maslach burnout survey" in the DevOps area, but missing sponsorship.
  • 27:00 - The Maslach burnout inventory is widely used in other industries.
  • 27:45 - There is a misunderstading that because IT salaries are high, IT staff has no worries.
  • 27:55 - Level of burnout in IT is probably equivalent to that of farmers and dockers.
  • 28:35 - Last AWS outage was blamed on a single person instead of looking at the system as a whole that led to that failure.
  • 30:01 - The new view on human error is that we need to look at the system that allowed people to make the mistakes in the first place.
  • 30:20 - The wide reach of software-controlled systems in everyday life has increased the potential for human errors to impact a lot of people.
  • 30:50 - This puts an unhealthy burden on the people running those systems.
  • 31:30 - Need to bring back the conversation around burnout.
  • 32:09 - Work/life balance is not as much about an individual, but the mismatch between organizational culture and the individual's stressers.
  • 32:45 - Maybe we need to change the discourse on mental health (often discredited in our industry) to focus on financial benefits of having more efficient people, by addressing burnout.

People mentioned

Tools mentioned

Conferences mentioned

Books mentioned

More about our podcasts

You can keep up-to-date with the podcasts via our RSS Feed, and they are available via SoundCloud, Apple Podcasts, Spotify, Overcast and the Google Podcast. From this page you also have access to our recorded show notes. They all have clickable links that will take you directly to that part of the audio.

Previous podcasts

Rate this Article

Adoption
Style

BT