The road to continuous deployment

Deploying code changes continuously to your production environment has many benefits. New features get to your users much faster which enables very short and quick iterations on them. And due to the small change sets you tend to encounter less bugs and regressions. In this talk I’ll present what Continuous Deployment means, and what you have to do, step-by-step, to make it work successfully for your project. We will cover topics like trunk based development, feature flags, partial roll outs, continuous integration, deployment automation and application monitoring.

Transcript

1. View Slide

2. What?
   

   View Slide

3. Why?
   

   View Slide

4. How?
   

   View Slide

5. A few words
   about me
   

   View Slide

6. View Slide

7. View Slide

8. http://speakerdeck.com/u/bastianhofmann
   

   View Slide

9. View Slide

10. How many
    releases?
    

    View Slide

11. Continuous
    Integration?
    

    View Slide

12. Continuous
    Deployment?
    

    View Slide

13. Feature flags?
    

    View Slide

14. Central log
    management?
    

    View Slide

15. KPIs?
    

    View Slide

16. If not, you will after
    the talk.
    

    View Slide

17. View Slide

18. So
    

    View Slide

19. Continuous
    Deployment
    

    View Slide

20. Deploying code
    changes to
    production
    continuously
    

    View Slide

21. Classic approach
    

    View Slide

22. Waterfall
    

    View Slide

23. http://en.wikipedia.org/wiki/File:Waterfall_model.svg
    

    View Slide

24. View Slide

25. Web! Hello?
    

    View Slide

26. Agile
    

    View Slide

27. http://en.wikipedia.org/wiki/File:Scrum_process.svg
    

    View Slide

28. Going further
    

    View Slide

29. Continuous
    Deployment
    

    View Slide

30. Benefits
    

    View Slide

31. Smaller change sets
    

    View Slide

32. Less Bugs
    

    View Slide

33. Finding the bug
    

    View Slide

34. Shortest „time to
    bugfix“ possible
    

    View Slide

35. Better performance
    

    View Slide

36. Better conversion
    

    View Slide

37. Faster iterations
    

    View Slide

38. Awesome, I’m sold
    

    View Slide

39. How can I
    introduce it?
    

    View Slide

40. The whole team
    has to live it
    

    View Slide

41. Not only doing
    more releases!
    

    View Slide

42. Step by step
    

    View Slide

43. is very
    important
    

    View Slide

44. Manual QA
    

    View Slide

45. does not scale
    

    View Slide

46. Automated Tests
    

    View Slide

47. Unit Tests
    

    View Slide

48. Integration Tests
    

    View Slide

49. Acceptance Tests
    

    View Slide

50. Smoke Tests
    

    View Slide

51. JS Tests
    

    View Slide

52. Regression Tests
    

    View Slide

53. Continuous
    Integration
    

    View Slide

54. Jenkins
    http://jenkins-ci.org/
    

    View Slide

55. View Slide

56. Development
    process
    

    View Slide

57. The problem with
    Feature Branches
    

    View Slide

58. View Slide

59. View Slide

60. Merge Hell
    

    View Slide

61. -> Big changes
    

    View Slide

62. Trunk/master based
    development
    

    View Slide

63. Every commit gets
    integrated
    

    View Slide

64. Every commit gets
    deployed
    

    View Slide

65. No commits that
    are not stable
    

    View Slide

66. Frequent release
    branches with
    cherry pics
    

    View Slide

67. http://paulhammant.com/2013/03/04/facebook-tbd/
    

    View Slide

68. http://paulhammant.com/2013/05/06/googles-scaled-trunk-based-development/
    

    View Slide

69. http://paulhammant.com/2013/03/04/facebook-tbd/
    

    View Slide

70. But there is stuff that
    takes longer then
    an hour to develop
    

    View Slide

71. Feature Flags/
    Toggles
    

    View Slide

72. Release !==
    Deployment
    

    View Slide

73. public function hasAccess(
    $accountId, array $roleNames
    ) {
    return featureFlag()->isActive(
    FeatureFlag::TEST_ONE
    );
    }
    

    View Slide

74. View Slide

75. View Slide

76. Partial rollout
    

    View Slide

77. Experiments (A/B Testing)
    

    View Slide

78. other also possible:
    goal early
    integration
    

    View Slide

79. Finally: Getting your
    code live
    

    View Slide

80. Automatic builds
    

    View Slide

81. Composer
    

    View Slide

82. Bower
    

    View Slide

83. NPM
    

    View Slide

84. Browserify
    

    View Slide

85. JS
    

    View Slide

86. CSS
    

    View Slide

87. HTML
    

    View Slide

88. Static caches
    

    View Slide

89. Ant
    

    View Slide

90. Phing
    

    View Slide

91. Grunt
    

    View Slide

92. View Slide

93. Custom scripts
    

    View Slide

94. Fast builds
    

    View Slide

95. Hook it into Jenkins
    

    View Slide

96. Automatic
    deployment
    

    View Slide

97. One-click-
    deployment
    

    View Slide

98. Fast deployment
    

    View Slide

99. Fast rollbacks
    

    View Slide

100. http://www.ansible.com/
     

     View Slide

101. http://capistranorb.com/
     

     View Slide

102. Etsy Deploynator
     http://codeascraft.com/2010/05/20/quantum-of-deployment/
     

     View Slide

103. Custom scripts
     

     View Slide

104. Zero downtime
     

     View Slide

105. Minimize the
     amount of time you
     have two different
     versions online in
     parallel
     

     View Slide

106. LB
     Green
     Blue
     

     View Slide

107. LB
     Green
     Blue
     Deploy new version
     

     View Slide

108. LB
     Green
     Blue
     

     View Slide

109. Attention: Path
     changes
     

     View Slide

110. Cache warmup
     

     View Slide

111. Asset invalidation
     

     View Slide

112. https://www.npmjs.org/package/grunt-cache-invalidate
     https://www.npmjs.org/package/gulp-rev
     https://www.npmjs.org/package/grunt-static-ctrl
     

     View Slide

113. Let’s deploy
     something
     

     View Slide

114. Staging System(s)
     

     View Slide

115. Central
     DEV
     Production
     Near
     Production
     Integration
     

     View Slide

116. Build/Test/Release
     pipeline
     

     View Slide

117. View Slide

118. Keep it fast
     

     View Slide

119. Allow manual
     shortcuts
     

     View Slide

120. Schema migrations
     

     View Slide

121. View Slide

122. Non BC breaking
     

     View Slide

123. Only safe
     operations
     

     View Slide

124. Migration in steps
     

     View Slide

125. No software is bug
     free
     

     View Slide

126. Finding bugs as
     early as possible
     

     View Slide

127. Detecting errors
     

     View Slide

128. Central log
     management
     

     View Slide

129. Logstash,
     Elasticsearch,
     Kibana
     

     View Slide

130. Detecting regressions
     

     View Slide

131. Bugs
     

     View Slide

132. Performance
     

     View Slide

133. KPIs
     

     View Slide

134. Plotting everything
     

     View Slide

135. Finding the cause
     

     View Slide

136. Dashboards
     

     View Slide

137. Anomaly detection
     

     View Slide

138. Alerts
     

     View Slide

139. Profiling
     

     View Slide

140. Synopsis
     

     View Slide

141. Invest in quality
     

     View Slide

142. Invest in
     performance
     

     View Slide

143. Invest in your
     processes
     

     View Slide

144. Invest in agility
     

     View Slide

145. Continuous
     Deployment
     

     View Slide

146. http://twitter.com/BastianHofmann
     http://lanyrd.com/people/BastianHofmann
     http://speakerdeck.com/u/bastianhofmann
     mail@bastianhofmann.de
     

     View Slide