Measuring Website Performance With New Relic

Slides from a hands-on workshop where attendees built a Node app and then monitored its performance using New Relic.

Transcript

1. CHICKTECH WORKSHOP 2016.05.18 MEASURING WEBSITE PERFORMANCE WITH NEW RELIC IAN DEES • @UNDEES
2. WHAT WE DO AT NEW RELIC
3. “UNDERSTAND YOUR
 DIGITAL BUSINESS”
4. FIND, FIX, AND PREVENT
 TECHNOLOGY PROBLEMS
 SO YOU CAN BUILD STUFF
5. APM APPLICATION PERFORMANCE MONITORING
6. WHY WE BUILT APM
7. TRF_Mr_Hyde
 https://flic.kr/p/bKumAD
 CC BY
8. <h1>My web page</h1> <p>The time is now.</p>
9. <h1>My web page</h1> <p>The time is <%= Time.now %>. </p>
10. <% $db = connect_to_database('db.example.com', 'password'); $rows = $db.query('select * from movies;'); %> <h1>My awesome movies</h1> <% for i in (0...$rows.count) do %> <p>Movie no. <%= i %> is <%= $rows[i][0] %></p> <% end %>
11. “I CAN’T KEEP TRACK OF ALL THIS
 TAG SOUP!” Web developers, ca. 2003
12. RAILS
13. class Movie < ActiveRecord::Base has_one :director has_many :performers end
14. class Movie < ActiveRecord::Base has_one :director has_many :performers end
15. “MY WEB APP IS SO SLOW!” Web developers, ca. 2007
16. App Server Database
17. APM
19. WHY APM IS IMPORTANT TODAY
20. APPS ARE EVEN MORE COMPLICATED NOW
21. Service Database Load Balancer App Server App Server
22. VARIETY OF LANGUAGES AND TECHNOLOGIES
23. WORKING ON THE BUSINESS >
 WORKING ON THE TECH
24. WHY APM IS IMPORTANT FOR YOU
25. YOU CAN SOLVE PROBLEMS MORE EFFECTIVELY
26. YOU CAN HIGHLIGHT THIS SKILL ON YOUR RESUME
28. YOU CAN HIGHLIGHT RESULTS ON YOUR RESUME
30. WE'RE HIRING, BY THE WAY HTTPS://NEWRELIC.COM/ABOUT/CAREERS
31. VARIETY OF EXPERIENCE
32. WANT TO SOLVE REAL PROBLEMS FOR PEOPLE
33. FAMILIARITY WITH APM IDEAS IS A PLUS
34. WE BUILD IT TOGETHER
35. WHAT WE'RE GOING TO DO IN THIS WORKSHOP
36. EVENTUALLY:
 A BLOG APP
37. WRITTEN IN NODE.JS
38. BACKED BY A DATABASE
39. WITH A FEW PERFORMANCE ISSUES
40. WE’RE GOING TO USE NEW RELIC TO FIND/FIX
41. TODAY: A STARTER “HELLO WORLD” APP
42. 1. GET THE TOOLS
43. 2. TEST THE APP
44. 3. ADD NEW RELIC
45. 4. SEE WHAT PROBLEMS LOOK LIKE
46. BACKGROUND CONCEPTS FIRST THINGS FIRST
47. FINDING SLOW QUERIES AN EASY WAY TO GET A QUICK WIN
48. DATABASE-DRIVEN WEB APPS
49. App Server Database
50. “CRUD” CREATE, READ, UPDATE, DESTROY
51. KEY IDEA: TABLES
52. EXAMPLE:
 MOVIE REVIEWS
53. HOW WOULD YOU TRACK MOVIES IN A SPREADSHEET?
55. Title Year Language OMG CAT 2005 Finnish LOL DOG 2010 Esperanto
56. App Server Database
57. SQL STRUCTURED QUERY LANGUAGE
58. Title Year Language OMG CAT 2005 Finnish LOL DOG 2010 Esperanto
59. FIND ALL MOVIES # SELECT * FROM movies; title | year | language ---------+------+----------- OMG CAT | 2005 | Finnish LOL DOG | 2010 | Esperanto (2 rows)
60. JUST THE TITLES # SELECT title FROM movies; title --------- OMG CAT LOL DOG (2 rows)
61. ONLY THE ONES IN ESPERANTO # SELECT * FROM movies WHERE language = 'Esperanto'; title | year | language ---------+------+----------- LOL DOG | 2010 | Esperanto (1 row)
62. MASHING UP DATA
63. MOVIES AND REVIEWS
64. IDEA #1:
 MOVIE TITLE IN THE REVIEW
65. Reviewer Stars Movie Siskeltron 4000 5 LOL DOG Ebert-bot 1 LOL DOG Ebert-bot 4 OMG CAT
66. WHAT KIND OF REVIEWS DID THIS MOVIE GET? # SELECT stars FROM reviews WHERE movie = 'LOL DOG'; stars ------- 5 1 (2 rows)
67. PROBLEM:
 WHAT HAPPENS WHEN
 WE CHANGE A TITLE?
68. Title Year Language OMG CAT 2005 Finnish LØL DØG 2010 Esperanto
69. Reviewer Stars Movie Siskeltron 4000 5 LOL DOG Ebert-bot 1 LOL DOG Ebert-bot 4 OMG CAT
70. IDEA #2:
 REVIEWS IN THE MOVIES
71. Title Year Language Stars OMG CAT 2005 Finnish 4 LOL DOG 2010 Esperanto 5,1
72. WHAT KIND OF REVIEWS DID THIS MOVIE GET? # SELECT stars FROM movies WHERE title = 'LOL DOG'; stars ------- 5,1 (1 row)
73. PROBLEM:
 WHERE'S THE REST
 OF THE REVIEW?
74. IDEA #3:
 REFER TO MOVIES
 BY NUMBER
75. ID Title Year Language 1 OMG CAT 2005 Finnish 2 LOL DOG 2010 Esperanto
76. Reviewer Stars Movie ID Siskeltron 4000 5 2 Ebert-bot 1 2 Ebert-bot 4 1
77. WHAT KIND OF REVIEWS DID THIS MOVIE GET? # SELECT * FROM reviews WHERE movie_id = 2; reviewer | stars | movie_id -----------------+-------+---------- Siskeltron 4000 | 5 | 2 Ebert-bot | 1 | 2 (2 rows)
78. AHA!
79. NORMALIZATION
80. HOW DO WE GET MOVIES AND REVIEWS TOGETHER?
81. STAY TUNED!
82. “N+1 PROBLEM”
83. BIG-O NOTATION ▸O(n)
 
 read as “O of n,” which stands for “on the order of n”
84. ORDERS OF COMPLEXITY
85. TRF_Mr_Hyde / https://flic.kr/p/9NuvP4 / CC BY
86. EXAMPLE:
 SORTING A LIST
87. HOW MANY STEPS?
88. BUBBLE SORT
90. ] 1
91. ] 1
92. ] 2
93. ] 2
94. ] 3
95. ] 3
97. ] 4
98. ] 4
99. ] 5
100. ] 5
102. ] 6
103. ] 6
105. Items Steps Total 4 3 + 2 + 1 6
106. Items Steps Total 2 1 1 3 2 + 1 3 4 3 + 2 + 1 6 5 4 + 3 + 2 + 1 10 6 5 + 4 + 3 + 2 + 1 15
107. Number of steps needed for n items = 1 + 2 + 3 + … + (n - 1) = ½ ⨉ n ⨉ (n − 1) = ½n2 − ½n

108. What order of complexity is this? ½n2 − ½n ½n2 − ½n ½n2 − ½n so, it's O(n2)
109. THESE ARE ALL THE SAME ORDER:
110. Time Items n2
111. Time Items n2 + 2
112. Time Items 2n2
113. Time Items n2 + n
114. BIG-O ANALYSIS ONE OF THE MOST POWERFUL TOOLS IN COMPUTING!
115. HELPS YOU PREDICT HOW YOUR PROGRAM IS GOING TO SCALE
116. DATABASE EXAMPLE
117. MOVIES AND REVIEWS
119. MOVIES AND REVIEWS # SELECT * FROM movies; id | title | year | language ----+---------+------+----------- 1 | OMG CAT | 2005 | Finnish 2 | LOL DOG | 2010 | Esperanto (2 rows) # SELECT * FROM reviews; reviewer | stars | movie_id -----------------+-------+---------- Siskeltron 4000 | 5 | 2 Ebert-bot | 1 | 2 Ebert-bot | 4 | 1 (3 rows)
120. THE HARD WAY:
 LOOP THROUGH MOVIES
121. MOVIES AND REVIEWS movies = db.query('SELECT * FROM movies;') movies.each do |movie| puts movie.title reviews = db.query('SELECT * FROM REVIEWS WHERE movie_id = ' + movie.id) reviews.each do |review| puts review.reviewer + ' gave it ' + review.stars + ' star(s)' end end
122. MOVIES AND REVIEWS movies = db.query('SELECT * FROM movies;') movies.each do |movie| puts movie.title reviews = db.query('SELECT * FROM REVIEWS WHERE movie_id = ' + movie.id) reviews.each do |review| puts review.reviewer + ' gave it ' + review.stars + ' star(s)' end end
123. MOVIES AND REVIEWS movies = db.query('SELECT * FROM movies;') movies.each do |movie| puts movie.title reviews = db.query('SELECT * FROM REVIEWS WHERE movie_id = ' + movie.id) reviews.each do |review| puts review.reviewer + ' gave it ' + review.stars + ' star(s)' end end
124. MOVIES AND REVIEWS movies = db.query('SELECT * FROM movies;') movies.each do |movie| puts movie.title reviews = db.query('SELECT * FROM REVIEWS WHERE movie_id = ' + movie.id) reviews.each do |review| puts review.reviewer + ' gave it ' + review.stars + ' star(s)' end end
125. MOVIES AND REVIEWS movies = db.query('SELECT * FROM movies;') movies.each do |movie| puts movie.title reviews = db.query('SELECT * FROM REVIEWS WHERE movie_id = ' + movie.id) reviews.each do |review| puts review.reviewer + ' gave it ' + review.stars + ' star(s)' end end
126. MOVIES AND REVIEWS movies = db.query('SELECT * FROM movies;') movies.each do |movie| puts movie.title reviews = db.query('SELECT * FROM REVIEWS WHERE movie_id = ' + movie.id) reviews.each do |review| puts review.reviewer + ' gave it ' + review.stars + ' star(s)' end end
127. Number of queries needed for n movies: = 1 query to get the list of movies + (n movies ⨉ 1 query for a single movie’s reviews) = n + 1

128. THINGS THAT HAPPEN FOR EVERY MOVIE: ▸ Your app builds a SQL query ▸ Your app sends the query over the network ▸ The database parses the query ▸ The database finds the answers on disk ▸ The database formats a SQL response ▸ The database sends the response over the network ▸ Your app decodes the response
129. N + 1 QUERIES TO LOAD THE PAGE 1000 movies ⨉ 1 review query for each movie = 1000 queries + 1 for the list of movie titles = 1001 queries
130. TIME FOR 1 PAGE LOAD Assume 0.01 second for a single query 1001 queries ⨉ 0.01 seconds for each query = 10.01 seconds to show your web page
131. WHAT IF WE COULD SEND JUST ONE BIG QUERY?
132. THINGS THAT HAPPEN FOR EVERY MOVIE JUST ONCE: ▸ Your app builds a SQL query ▸ Your app sends the query over the network ▸ The database parses the query ▸ The database finds the answers on disk ▸ The database formats a SQL response ▸ The database sends the response over the network ▸ Your app decodes the response
133. THESE BITS ARE SLOWER: ▸ Your app builds a SQL query ▸ Your app sends the query over the network ▸ The database parses the query ▸ The database finds the answers on disk ▸ The database formats a SQL response ▸ The database sends the response over the network ▸ Your app decodes the response
134. SO LET’S SAY A BIG QUERY TAKES 100X LONGER
135. (IT WON’T, BUT LET’S BE PESSIMISTS)
136. TIME FOR 1 PAGE LOAD (WITH ONE BIG QUERY) Assume the “give me everything” query takes: 100 times longer ⨉ 0.01 seconds = 1 second to show your web page …a 10x speedup!
137. HOW TO WRITE
 ONE BIG QUERY?
138. SQL JOINS THE DETAIL BEHIND THE SCENES
139. MOVIES AND REVIEWS # SELECT * FROM movies; id | title | year | language ----+---------+------+----------- 1 | OMG CAT | 2005 | Finnish 2 | LOL DOG | 2010 | Esperanto (2 rows) # SELECT * FROM reviews; reviewer | stars | movie_id -----------------+-------+---------- Siskeltron 4000 | 5 | 2 Ebert-bot | 1 | 2 Ebert-bot | 4 | 1 (3 rows)
140. WE WANT A MASHUP:
141. Reviewer Stars Title Siskeltron 4000 5 LOL DOG Ebert-bot 1 LOL DOG Ebert-bot 4 OMG CAT
142. JOIN
143. MOVIES THAT HAVE REVIEWS # SELECT reviews.reviewer, reviews.stars, movies.title FROM reviews INNER JOIN movies ON (reviews.movie_id = movies.id); reviewer | stars | title -----------------+-------+--------- Siskeltron 4000 | 5 | LOL DOG Ebert-bot | 1 | LOL DOG Ebert-bot | 4 | OMG CAT (3 rows)
144. MOVIES THAT HAVE REVIEWS # SELECT reviews.reviewer, reviews.stars, movies.title FROM reviews INNER JOIN movies ON (reviews.movie_id = movies.id); reviewer | stars | title -----------------+-------+--------- Siskeltron 4000 | 5 | LOL DOG Ebert-bot | 1 | LOL DOG Ebert-bot | 4 | OMG CAT (3 rows)
145. MOVIES THAT HAVE REVIEWS # SELECT reviews.reviewer, reviews.stars, movies.title FROM reviews INNER JOIN movies ON (reviews.movie_id = movies.id); reviewer | stars | title -----------------+-------+--------- Siskeltron 4000 | 5 | LOL DOG Ebert-bot | 1 | LOL DOG Ebert-bot | 4 | OMG CAT (3 rows)
146. MOVIES THAT HAVE REVIEWS # SELECT reviews.reviewer, reviews.stars, movies.title FROM reviews INNER JOIN movies ON (reviews.movie_id = movies.id); reviewer | stars | title -----------------+-------+--------- Siskeltron 4000 | 5 | LOL DOG Ebert-bot | 1 | LOL DOG Ebert-bot | 4 | OMG CAT (3 rows)
147. MOVIES THAT HAVE REVIEWS # SELECT reviews.reviewer, reviews.stars, movies.title FROM reviews INNER JOIN movies ON (reviews.movie_id = movies.id); reviewer | stars | title -----------------+-------+--------- Siskeltron 4000 | 5 | LOL DOG Ebert-bot | 1 | LOL DOG Ebert-bot | 4 | OMG CAT (3 rows)
148. YOUR APP FRAMEWORK WILL HAVE SHORTCUTS
149. RAILS WITH AN N+1 class Movie < ActiveRecord::Base has_many reviews end movies = Movie.all # ...then, get reviews for each movie...
150. N+1 FIXED class Movie < ActiveRecord::Base has_many reviews end movies = Movie.includes(:reviews) # ...then, get reviews for each movie...
151. FIXING N+1 CAN BE YOUR SUPERPOWER
152. NEXT TIME, YOU’LL DO JUST THAT
153. TIME FOR A BREAK!
154. LEGALESE This document and the information herein (including any information that may be incorporated by reference) is provided for informational purposes only and should not be construed as an offer, commitment, promise or obligation on behalf of New Relic, Inc. (“New Relic”) to sell securities or deliver any product, material, code, functionality, or other feature. Any information provided hereby is proprietary to New Relic and may not be replicated or disclosed without New Relic’s express written permission. Such information may contain forward-looking statements within the meaning of federal securities laws. Any statement that is not a historical fact or refers to expectations, projections, future plans, objectives, estimates, goals, or other characterizations of future events is a forward-looking statement. These forward-looking statements can often be identified as such because the context of the statement will include words such as “believes,” “anticipates,” “expects” or words of similar import. Actual results may differ materially from those expressed in these forward-looking statements, which speak only as of the date hereof, and are subject to change at any time without notice. Existing and prospective investors, customers and other third parties transacting business with New Relic are cautioned not to place undue reliance on this forward-looking information. The achievement or success of the matters covered by such forward-looking statements are based on New Relic’s current assumptions, expectations, and beliefs and are subject to substantial risks, uncertainties, assumptions, and changes in circumstances that may cause the actual results, performance, or achievements to differ materially from those expressed or implied in any forward-looking statement. Further information on factors that could affect such forward-looking statements is included in the filings we make with the SEC from time to time. Copies of these documents may be obtained by visiting New Relic’s Investor Relations website at ir.newrelic.com or the SEC’s website at www.sec.gov. New Relic assumes no obligation and does not intend to update these forward-looking statements, except as required by law. New Relic makes no warranties, expressed or implied, in this document or otherwise, with respect to the information provided.