deep devops - learning to learn dotScale 2017

Post on 22-Jan-2018

491 views 0 download

transcript

deep devops learning to learn

Andrew Clay Shafer @littleidea

Andrew Clay Shafer

Andrew Clay Shafer

@littleidea

Andrew Clay Shafer

@littleidea

@littleidea

Agile Infrastructure

@littleidea’sgreatest hits compilation

Change Management

There Is No Talent Shortage

leading a learning organization

everyone wants the devops

where are we? how did we get here?

Operations is the secret sauce

you can either easily manage complex systems at scale…

or you can’t

I’ve been using that same slide for the better part of a decade

fascinated with the dynamics of high performing systems and teams

same tools, same advice, drastically different results

what, how and why you automate is as important that you do

yay automation!

I’m sure some of you have lived this too

or are about to

yay automation!

If Tetris Has Taught Me Anything, It’s That Errors Pile Up and

Accomplishments Disappear

why > what

Pareto Inefficient Nash Equilibria Rule Everything Around You

Pareto Inefficient - a change benefits someone without detriment to anyone

Nash Equilibrium - no one will change

at a Nash Equilibrium there is no benefit to changing

true by definition

Punctuated Equilibriumevolutionary gradualism is virtually non-existent in the fossil record

Evolution in the fossil record comes in sudden jumps

and extinction events

Nash Equilibria til death do us part

lose if you keep playing the same way

what is the devops?

devops• developers and operations can and

should work together

• system administration evolving to look more like software development

• evolving together as global community sharing solutions

devops - calms• culture

• automation

• lean

• metrics

• sharing the five pillars

optimizing human performance and experience operating software…

and humans

with software…

software is changing everything

including software

and organizations

you are either building a software business...

or you will be losing to someone who is...

you are either building a learning organization...

or you will be losing to someone who is...

everyone wants the devops

Well actually…

what they really want• scalability

• availability

• reliability

• operability

• usability

• all for free

• without changing anything

good luck; have fun

devops, continuous delivery, microservices…

These things are all one…

continuously devops microservices

or die trying

the patterns proven successful building and operating highly

available systems with predictable scaling and failure characteristics

devops as she is spoken at scaleat Google™

not that Google is the best at every possible thing, but where is your book?

SRE• developers and operations

can and should work together

• solving system administration with software development

• participating in the global community sharing solutions

SRE - calms• culture

• automation

• lean

• metrics

• sharing

Homework• Embracing Risk • Service Level Objectives • Eliminating Toil

Bonus: Communication and Collaboration in SRE

why > what

Google is an organization that changes

who puts as much effort into designing organizations as our computational systems?

how many of our organizations are able and willing to change?

how many of us are willing and able to change ourselves?

You haven’t learned anything until you change your behavior

The learning organization is one that has the capacity to integrate people and structures in order to move toward continuous learning and change.

no one originally set out to do devops, continuous delivery, microservices, these were natural consequences

don’t fixate on the words, don’t fixate on the tools, fixate on the outcomes

principles > practices > tools

mindset > skillset > toolset

adapt > adopt

The problem isn’t technical. The problem isn’t people.

The problem is socio-technical.We have to solve both together.

It is not the strongest of the species that survives, nor the most intelligent that survives. It is the one that is most adaptable to change.

— No evidence Darwin ever said this

what happens next?

change

my next talk: ’the science of why you shouldn’t

change everything at once’

@littleidea

I don’t hope to have answers I hope to ask better questions

@littleidea

We are uncovering better ways of developing software, by doing it and helping others do it