Slide 1

Slide 1 text

Not Everyone Can Be Kelsey Hightower Greg Poirier Factotum @ Opsee @grepory

Slide 2

Slide 2 text

Who am I?

Slide 3

Slide 3 text

What is Opsee?

Slide 4

Slide 4 text

What does the first engineer do, anyway? What even is a Factotum?

Slide 5

Slide 5 text

No content

Slide 6

Slide 6 text

Proof of Concept

Slide 7

Slide 7 text

Mini mum Viabl e Prod uct

Slide 8

Slide 8 text

I needed infrastructure.

Slide 9

Slide 9 text

EFPR Loop

Slide 10

Slide 10 text

Experiment Fail Process Rebuild

Slide 11

Slide 11 text

Bloom’s Taxonomy of the Cognitive Domain

Slide 12

Slide 12 text

Knowledge Comprehension Analysis Application Synthesis Evaluation

Slide 13

Slide 13 text

Knowledge

Slide 14

Slide 14 text

No content

Slide 15

Slide 15 text

Comprehension

Slide 16

Slide 16 text

Why do we need more than screen?

Slide 17

Slide 17 text

No content

Slide 18

Slide 18 text

Application

Slide 19

Slide 19 text

No content

Slide 20

Slide 20 text

Analysis

Slide 21

Slide 21 text

No content

Slide 22

Slide 22 text

No content

Slide 23

Slide 23 text

Green fielding sure is fun!

Slide 24

Slide 24 text

No content

Slide 25

Slide 25 text

Let's rub some Kubernetes on it.

Slide 26

Slide 26 text

And this is when we come to Kelsey Hightower.

Slide 27

Slide 27 text

Who is Kelsey Hightower?

Slide 28

Slide 28 text

Why do we want to be Kelsey Hightower?

Slide 29

Slide 29 text

No content

Slide 30

Slide 30 text

Why can’t we all just be Kelsey Hightower?

Slide 31

Slide 31 text

No content

Slide 32

Slide 32 text

Kubernetes is really amazing—if you know what you’re doing.

Slide 33

Slide 33 text

CORROLLARY: If you don’t know what you’re doing, Kubernetes is overwhelming.

Slide 34

Slide 34 text

Synthesis

Slide 35

Slide 35 text

No content

Slide 36

Slide 36 text

When considering what to do, optimize for maximum long-term gain.

Slide 37

Slide 37 text

Repeatable automated builds.

Slide 38

Slide 38 text

quay.io/opsee/build-go quay.io/opsee/build-clj

Slide 39

Slide 39 text

gliderlabs/alpine

Slide 40

Slide 40 text

No content

Slide 41

Slide 41 text

Deploy by restarting.

Slide 42

Slide 42 text

fleetctl stop foo@{1..n}.service fleetctl start foo@{1..n}.service

Slide 43

Slide 43 text

docker run quay.io/ opsee/service:latest

Slide 44

Slide 44 text

docker run quay.io/ opsee/service:${TIER}

Slide 45

Slide 45 text

So it was totally easy, right?

Slide 46

Slide 46 text

Everything is difficult.

Slide 47

Slide 47 text

Ignorance begets frustration.

Slide 48

Slide 48 text

Enter etcd.

Slide 49

Slide 49 text

STORY TIME

Slide 50

Slide 50 text

This one time… at CoreOS camp…

Slide 51

Slide 51 text

Learn etcd the hard way.

Slide 52

Slide 52 text

No content

Slide 53

Slide 53 text

Evaluation

Slide 54

Slide 54 text

This is not so bad.

Slide 55

Slide 55 text

No content

Slide 56

Slide 56 text

No content

Slide 57

Slide 57 text

No content

Slide 58

Slide 58 text

“My biggest obstacle was understanding CoreOS and reconciling some faulty assumptions, but I kind of put that off until last week.”

Slide 59

Slide 59 text

No content

Slide 60

Slide 60 text

Just say no to copy-and-paste infrastructure.

Slide 61

Slide 61 text

CoreOS is a Linux container hypervisor… not a minor deity.

Slide 62

Slide 62 text

CoreOS+Docker+Fleet = Tons of infrastructure code we didn't have to write.

Slide 63

Slide 63 text

We need a toolbox. quay.io/opsee/toolbox

Slide 64

Slide 64 text

That’s the story of how Opsee infrastructures and how I <3 Kelsey Hightower.

Slide 65

Slide 65 text

Thanks! @grepory - @opseeco opsee.com