Slide 1

Slide 1 text

Lies, Damn Lies, and Metrics

Slide 2

Slide 2 text

André Arko @indirect

Slide 3

Slide 3 text

No content

Slide 4

Slide 4 text

Bundler Managing application dependencies since 2009

Slide 5

Slide 5 text

Metrics

Slide 6

Slide 6 text

Metrics are important

Slide 7

Slide 7 text

Metrics tell you what is happening

Slide 8

Slide 8 text

you rn →

Slide 9

Slide 9 text

Metrics convince you you understand

Slide 10

Slide 10 text

you later →

Slide 11

Slide 11 text

Averages convince you you understand

Slide 12

Slide 12 text

Averages are lie-candy for your brain

Slide 13

Slide 13 text

“Normal” 5 -5 -4 -3 -2 -1 0 1 2 3 4 0 0.1 0.2 0.3 0.4

Slide 14

Slide 14 text

“Normal” 5 -5 -4 -3 -2 -1 0 1 2 3 4 0 0.1 0.2 0.3 0.4

Slide 15

Slide 15 text

Real Life 5 -5 -4 -3 -2 -1 0 1 2 3 4 0 0.1 0.2 0.3 0.4

Slide 16

Slide 16 text

brendangregg.com

Slide 17

Slide 17 text

brendangregg.com

Slide 18

Slide 18 text

just heard “w e have a great average” →

Slide 19

Slide 19 text

The problem with averages: If you put one hand in a bucket of ice and the other in a bucket of hot coals, on average, you’re comfortable. Erik Michaels-Ober @sferik

Slide 20

Slide 20 text

Averages mask problems

Slide 21

Slide 21 text

10 0 1 2 3 4 5 6 7 8 9 250 0 50 100 150 200

Slide 22

Slide 22 text

Graph the median

Slide 23

Slide 23 text

10 0 1 2 3 4 5 6 7 8 9 250 0 50 100 150 200

Slide 24

Slide 24 text

Graph 95th percentile

Slide 25

Slide 25 text

10 0 1 2 3 4 5 6 7 8 9 250 0 50 100 150 200

Slide 26

Slide 26 text

Graph 99th percentile

Slide 27

Slide 27 text

10 0 1 2 3 4 5 6 7 8 9 1000 0 250 500 750

Slide 28

Slide 28 text

Aggregate graphs another average

Slide 29

Slide 29 text

No content

Slide 30

Slide 30 text

Breakout graphs show each source

Slide 31

Slide 31 text

No content

Slide 32

Slide 32 text

Seriously, do it Visualize your data

Slide 33

Slide 33 text

graphic by Schutz and Avenue, CC-Attribution-ShareAlike, taken from from the Wikipedia article on Anscombe's quartet

Slide 34

Slide 34 text

Average of X: 9 Average of X: 9 Average of X: 9 Average of X: 9

Slide 35

Slide 35 text

Average of Y: 7.50 Average of Y: 7.50 Average of Y: 7.50 Average of Y: 7.50

Slide 36

Slide 36 text

Average of X Average of Y Variance of X Variance of Y Correlation of X and Y Linear regression All four data sets have the same

Slide 37

Slide 37 text

Aggregate alerts more dead servers than alive servers

Slide 38

Slide 38 text

site’s up if any servers are up!

Slide 39

Slide 39 text

Breakout alerts first dead server not all the servers

Slide 40

Slide 40 text

Servers

Slide 41

Slide 41 text

Servers you have no idea what is going on

Slide 42

Slide 42 text

really.

Slide 43

Slide 43 text

Runtime lag

Slide 44

Slide 44 text

Runtime lag how do you tell you lost consciousness?

Slide 45

Slide 45 text

Runtime lag you have it.

Slide 46

Slide 46 text

Runtime lag you have it. how bad is it?

Slide 47

Slide 47 text

VM lag

Slide 48

Slide 48 text

VM lag do you have it?

Slide 49

Slide 49 text

VM lag do you check for it?

Slide 50

Slide 50 text

VM lag do you know how to check for it?

Slide 51

Slide 51 text

Routing

Slide 52

Slide 52 text

Routing your app has this

Slide 53

Slide 53 text

Routing how does it work?

Slide 54

Slide 54 text

Development App You

Slide 55

Slide 55 text

Production People Router Server App App Router Server App App Router

Slide 56

Slide 56 text

Routing how slow is it?

Slide 57

Slide 57 text

Routing does it back up?

Slide 58

Slide 58 text

Request time

Slide 59

Slide 59 text

Request time not the time you measure

Slide 60

Slide 60 text

Request time wall-clock time from real clients

Slide 61

Slide 61 text

Request time make requests from around the world

Slide 62

Slide 62 text

metrics are good So, in the end

Slide 63

Slide 63 text

know what you are measuring but

Slide 64

Slide 64 text

@indirect [email protected] Questions?