DevOps Theory vs. Practice: A Song of Ice and TireFire

A presentation at Austin DevOps Meetup in March 2019 in Austin, TX, USA by Baruch Sadogursky

Slide 1

Slide 1

Software Engineering Theory vs. Practice: A Song of Ice and Tire Fire

Slide 2

Slide 2

Senior SRE Leader at Google Thought Leader, Disruptive Innovator Senior Software Engineer at Netflix SVP of Thoughts at Facebook Obviously better than you Disclaimer: absolutely

Slide 3

Slide 3

Official Hiptech Translator Native proficiency in Russian, Hebrew, English, Java, Go Curses in 18 more languages Fluent in Thought Leader gibberish Disclaimer: some of the above may or may not be true. @jbaruch @austindevops jfrog.com/shownotes

Slide 4

Slide 4

Baruch, Thought Leader Away! @jbaruch @austindevops jfrog.com/shownotes

Slide 5

Slide 5

Everybody’s software must be releasable at absolutely any time @jbaruch @austindevops jfrog.com/shownotes

Slide 6

Slide 6

Everyone must have 100% test automation @jbaruch @austindevops jfrog.com/shownotes

Slide 7

Slide 7

We do Continuous Security well. @jbaruch @austindevops jfrog.com/shownotes

Slide 8

Slide 8

Your greatest threat is an outage. Not an employee. @jbaruch @austindevops jfrog.com/shownotes

Slide 9

Slide 9

VMs are the enemy of DevOps. This is where you must focus your innovation. @jbaruch @austindevops jfrog.com/shownotes

Slide 10

Slide 10

You are a beautiful unique snowflake, as are your problems. No vendor could possibly understand them. @jbaruch @austindevops jfrog.com/shownotes

Slide 11

Slide 11

Our company is based in SF because that’s where the best engineers are. @jbaruch @austindevops jfrog.com/shownotes

Slide 12

Slide 12

THANK YOU! @jbaruch @austindevops jfrog.com/shownotes

Slide 13

Slide 13

Slide 14

Slide 14

Shownotes! ● ● ● ● ● ● jfrog.com/shownotes Slides Video (by tomorrow) All the links! Comments, Ratings Raffle! @jbaruch @austindevops jfrog.com/shownotes

Slide 15

Slide 15

How did we get here? @jbaruch @austindevops jfrog.com/shownotes

Slide 16

Slide 16

Slide 17

Slide 17

TIMEThereFOR BULLET POINTS could be text here too • Gotta love bullets • Gotta love bullets • Gotta love bullets • Gotta love bullets • Gotta love bullets • Gotta love bullets

Slide 18

Slide 18

Slide 19

Slide 19

TIMEThereFOR BULLET POINTS could be text here too • Gotta love bullets • Gotta love bullets • Gotta love bullets • Gotta love bullets • Gotta love bullets • Gotta love bullets

Slide 20

Slide 20

Slide 21

Slide 21

Cargo Cult @jbaruch @austindevops jfrog.com/shownotes

Slide 22

Slide 22

The Four Questions @jbaruch @austindevops jfrog.com/shownotes

Slide 23

Slide 23

The Four Questions 1. Is my organization/team ready to adopt a new tech? 2. Is it even a good tech? 3. What do I gain from adopting this tech? 4. Is this tech a good solution to my problem?

Slide 24

Slide 24

The Four Questions 1. Is my organization/team ready to adopt a new tech? 2. Is it even a good tech? 3. What do I gain from adopting this tech? 4. Is this tech a good solution to my problem?

Slide 25

Slide 25

Is my organization /team ready to adopt a new tech? 1. @jbaruch @austindevops jfrog.com/shownotes

Slide 26

Slide 26

• “A maturity model is a tool that helps people assess the current effectiveness of a person or group and supports figuring out what capabilities they need to acquire next in order to improve their performance. Introducing maturity models • In many circles maturity models have gained a bad reputation, but although they can easily be misused, in proper hands they can be helpful.” • Martin Fowler @jbaruch @austindevops jfrog.com/shownotes

Slide 27

Slide 27

• “A maturity model is a tool that helps people assess the current effectiveness of a person or group and supports figuring out what capabilities they need to acquire next in order to improve their performance. Introducing maturity models • In many circles maturity models have gained a bad reputation, but although they can easily be misused, in proper hands they can be helpful.” • Martin Fowler @jbaruch @austindevops jfrog.com/shownotes

Slide 28

Slide 28

Introducing maturity models Nicole Forsgren, Jezz Hamble, Gene Kim @jbaruch @austindevops jfrog.com/shownotes

Slide 29

Slide 29

Introducing maturity models Nicole Forsgren, Jezz Hamble, Gene Kim @jbaruch @austindevops jfrog.com/shownotes

Slide 30

Slide 30

Bad Maturity Models are Bad. Maturity is a process, not a goal One size does not fit all organizations (or teams in organization) Focus on outcomes, not existence of a process or a tool Always evolve the model as the industry and the organization evolves

Slide 31

Slide 31

Maturity model components Evaluation factors Scoring methodology Self assessment vs 3rd party assessment capability Progress tracking Visualization @jbaruch @austindevops jfrog.com/shownotes

Slide 32

Slide 32

Maturity Model Examples @jbaruch @austindevops jfrog.com/shownotes

Slide 33

Slide 33

Simple model @jbaruch @austindevops jfrog.com/shownotes

Slide 34

Slide 34

Progress planning Target Today @jbaruch @austindevops In 2Q jfrog.com/shownotes

Slide 35

Slide 35

Leader board Team B Team A Team C Team D Team E @jbaruch @austindevops jfrog.com/shownotes

Slide 36

Slide 36

@jbaruch @austindevops jfrog.com/shownotes

Slide 37

Slide 37

@jbaruch @austindevops jfrog.com/shownotes

Slide 38

Slide 38

A different approach @jbaruch @austindevops jfrog.com/shownotes

Slide 39

Slide 39

@jbaruch @austindevops jfrog.com/shownotes

Slide 40

Slide 40

Account for different teams’ priorities @jbaruch @austindevops jfrog.com/shownotes

Slide 41

Slide 41

Model definition example System config as Code The infrastructure configuration is managed as code - e.g. no manual processes for configuring/setting up/ infrastructure. Differentiating: Infrastructure operates without any manual processes. All changes to the infrastructure or infrastructure capabilities are done through automation and policy only. Complete: Infrastructure operates without any manual processes. Some infrequent administrative activities may be initiated manually (although the activities themselves must be automated). Partial (Most): Infrastructure operates without any manual processes. Some infrequent administrative activities may be manual, pending automation. Partial (Much): Infrastructure operates with significant automation. Some processes still manual; pending automation. Partial (Some): Infrastructure requires significant care and feeding. Many processes still manual; pending automation. No Support: While some functions may be automated, they are generally kicked-off manually; and many functions are still fully manual. Large backlog of automation items. @jbaruch @austindevops jfrog.com/shownotes

Slide 42

Slide 42

Applying maturity models: DOs and DONT’s Only use primary colors Involve your teams in the model definition Let team self assess first and then assess together Partner with forward looking teams first Remember being at 100% is not a goal the model has to have a stretch goal Evolve the model from time to time And …. @jbaruch @austindevops jfrog.com/shownotes

Slide 43

Slide 43

Our message is: https://martinfowler.com/bliki/MaturityModel.html @jbaruch @austindevops jfrog.com/shownotes

Slide 44

Slide 44

@jbaruch Is it even a good tech? @austindevops jfrog.com/shownotes

Slide 45

Slide 45

@jbaruch @austindevops jfrog.com/shownotes

Slide 46

Slide 46

@jbaruch @austindevops jfrog.com/shownotes

Slide 47

Slide 47

@jbaruch @austindevops jfrog.com/shownotes

Slide 48

Slide 48

@jbaruch @austindevops jfrog.com/shownotes

Slide 49

Slide 49

The Recap: The Four Questions 1. Is my organization/team ready to adopt a new tech? 2. Is it even a good tech? 3. What do I gain from adopting this tech? 4. Is this tech a good solution to my problem?

Slide 50

Slide 50

Thank you very much! ● ● ● Shownotes! @jbaruch @austindevops @jbaruch @austindevops jfrog.com/shownotes

Slide 51

Slide 51

Things that are not questions ● ● ● Your résumé Calling bullshit on the entire premise of the talk A long rambling story with no point @jbaruch @austindevops jfrog.com/shownotes