Real Life
5
-5 -4 -3 -2 -1 0 1 2 3 4
0
0.1
0.2
0.3
0.4
Slide 40
Slide 40 text
brendangregg.com
Slide 41
Slide 41 text
brendangregg.com
Slide 42
Slide 42 text
just heard
“w
e
have
a
great average” →
Slide 43
Slide 43 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 44
Slide 44 text
Averages
mask problems
Slide 45
Slide 45 text
10
0 1 2 3 4 5 6 7 8 9
250
0
50
100
150
200
Slide 46
Slide 46 text
Graph
the median
Slide 47
Slide 47 text
10
0 1 2 3 4 5 6 7 8 9
250
0
50
100
150
200
Slide 48
Slide 48 text
Graph
95th percentile
Slide 49
Slide 49 text
10
0 1 2 3 4 5 6 7 8 9
250
0
50
100
150
200
Slide 50
Slide 50 text
Graph
99th percentile
Slide 51
Slide 51 text
10
0 1 2 3 4 5 6 7 8 9
1000
0
250
500
750
Slide 52
Slide 52 text
Aggregate graphs
another average
Slide 53
Slide 53 text
No content
Slide 54
Slide 54 text
Breakout graphs
show each source
Slide 55
Slide 55 text
No content
Slide 56
Slide 56 text
Seriously, do it
Visualize your data
Slide 57
Slide 57 text
graphic by Schutz and Avenue, CC-Attribution-ShareAlike, taken from from the Wikipedia article on Anscombe's quartet
Slide 58
Slide 58 text
Average of X: 9 Average of X: 9
Average of X: 9 Average of X: 9
Slide 59
Slide 59 text
Average of Y: 7.50 Average of Y: 7.50
Average of Y: 7.50 Average of Y: 7.50
Slide 60
Slide 60 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 61
Slide 61 text
Aggregate alerts
more dead servers
than alive servers
Slide 62
Slide 62 text
site’s up if any
servers are up!
Slide 63
Slide 63 text
Breakout alerts
first dead server
not all the servers
Slide 64
Slide 64 text
So, in the end
Slide 65
Slide 65 text
don’t forget
the network
Slide 66
Slide 66 text
user experience
is all that matters
after deploying