Fast Deployment and Monitoring For HelsinkiJS

Talk given at HelsinkiJS. Updated slides and colors for presentation

Transcript

1. Fast Deployment and Monitoring Why continuous integration and ChatOps are crucial to every application
2. Agenda Who am I? Testing Hooks / Services ChapOps
3. Who am I?
4. Trainer
5. Developer
6. In San Francisco now
8. Who even writes tests?
9. Who even writes tests first ?
10. I never did while in school
11. I still don't
12. I still don't
13. I'm geing beer!
14. Tests
15. "They take too long to write!"
16. "They take too long to write!" "I don't know how to write them"
17. "They take too long to write!" "I don't know how to write them" "My code is perfect"
18. "They take too long to write!" "I don't know how to write them" "My code is perfect" Lies! Lies! Lies!
19. Make it easy to run tests
20. script/cibuild
23. Sometimes running tests takes too long
24. Sometimes running tests takes too long
25. "Well older things still pass tests"
26. Share burden and responsibility
27. Never work alone
28. Never work alone
29. Pull Requests
30. Send yourself Pull Requests
31. Get others to do the work for you!
32. @mentions
33. @mentions
34. What if others could see if your tests were passing without running the code?
35. Hooks and Services
36. Make tests run easily for services
37. script/cibuild
38. Many third party services can be hooked into GitHub
39. Status API
40. Now someone visiting the Pull Request can know what's going on
41. Status Types
42. History of timeline statuses
43. Details
45. So many CI services
49. Self Configuration
52. Manual Configuration
56. Services can do more than just send data to the API
57. ChatOps
58. Context of the current state of things
59. Why Chat?
60. Hubot
61. Hubot
62. hubot.github.com
67. Hubot's Domain
68. Who does the things? Service? Hook? Robot?
69. Hubot: deploy all the things
70. Make it fast
71. Hubot: Make it fast
72. Deploys
75. "Hello @user, your deploy was successful, keep an eye on Haystack"
76. "Hello @user, your deploy was successful, keep an eye on Haystack"
78. Digging Deeper
79. Going beyond an exception
80. Graphite
81. Graphite
82. Now we have things reporting
83. Now we have things reporting
84. Who should fetch these graphs?
88. Maybe just a dashboard view?
89. Dashing
91. Dashboards
93. Group understanding
94. Group understanding
95. Useful commands
96. Hubot: help pager
97. Hubot: graph me -12h @network.border.cp1.bgp
98. Hubot: graph me -4h @browser.frontend
99. Hubot: deploy training- app/chronological to production
100. Teach by doing
101. This was always my main motivation with Hubot - teaching by doing by making this visible. It's an extremely powerful teaching technique. - @rtomayko
102. Signal vs Noise
103. Next Steps
104. Easy to run tests
105. More tests get wrien
106. Or collaborate
107. What's the state of the Pull Request?
108. What has been happening on the Pull Request
109. Your chatroom could be your office
110. Use robots
111. Use robots more
112. How do things look?
113. Don't bother me
114. Keep Deploying. Keep Building. Do it Faster & Safer.
115. Thank you! Brent Beer GitHub Trainer @brntbeer hp://bit.ly/fast-deployment-monitoring-helsinkijs
116. References: hps://hubot.github.com hps://github.com/shopify/dashing hps://github.com/graphite-project hps://github.com/etsy/statsd/ hps://developer.github.com/v3/ hps://github.com/collectd hps://developer.github.com/v3/ hps://circleci.com/ hps://travis-ci.org/ hps://github.com/github/janky