Deathmatch: Bundler vs. Rubygems (GoRuCo 2013)

The story of the quest to make bundle install faster; in which Rubyists around the world inadvertently DDoS rubygems.org, witness its ignominious death, and vow to rebuild it from the ashes stronger than it was before. Then, a tour of the changes; why is Redis so much slower than Postgres? Marvel at the gorgeous metrics and graphs used to measure and optimize; gasp in delight as we track, live, exactly how many Heroku dynos are needed. Finally, a happy ending: today, the server responds to requests TWO ORDERS OF MAGNITUDE faster than it did before.

Transcript

1. Deathmatch
   vs
   

   View Slide

2. André Arko
   @indirect
    
    
   

   View Slide

3. View Slide

4. View Slide

5. vs
   

   View Slide

6. !
   

   View Slide

7. Bundler 1.0
   

   View Slide

8. Bundler 1.0
   30 Aug 2010
   

   View Slide

9. Bundler 1.0
   30 Aug 2010
   !
   

   View Slide

10. bundle install
    

    View Slide

11. bundle install
    ...and install
    

    View Slide

12. bundle install
    ...and install
    ...and install
    

    View Slide

13. bundle install
    ...and install
    ...and install
    ...and install
    ...and install
    

    View Slide

14. !
    

    View Slide

15. (what seemed like)
    10 minutes later…
    

    View Slide

16. View Slide

17. bundler API
    

    View Slide

18. just the gems
    you need
    bundler API
    

    View Slide

19. so. much. faster.
    bundler API
    

    View Slide

20. so. much. faster.
    32s → 8s
    

    View Slide

21. Bundler 1.1
    

    View Slide

22. 07 March 2012
    Bundler 1.1
    

    View Slide

23. !
    07 March 2012
    Bundler 1.1
    

    View Slide

24. are lots of work
    API requests
    

    View Slide

25. etc, etc…
    pride cometh
    

    View Slide

26. reenactment
    dramatic
    

    View Slide

27. !
    !
    

    View Slide

28. !
    !
    !
    

    View Slide

29. !
    !
    !
    ! !
    

    View Slide

30. !
    !
    !
    ! !
    ! !
    

    View Slide

31. no more API
    

    View Slide

32. View Slide

33. View Slide

34. View Slide

35. crap.
    

    View Slide

36. replacement
    a new bundler api
    

    View Slide

37. heroku
    sinatra
    postgres
    

    View Slide

38. questions
    

    View Slide

39. questions
    will it work,
    this time?
    

    View Slide

40. questions
    how is it bad?
    

    View Slide

41. questions
    how can it be good?
    

    View Slide

42. answers
    need measurements
    

    View Slide

43. we measure
    a lot of things
    

    View Slide

44. librato
    tracks
    all the metrics
    

    View Slide

45. View Slide

46. papertrail
    logs
    all the lines
    

    View Slide

47. View Slide

48. knowledge
    measuring leads to
    

    View Slide

49. discovery 1
    postgres is fast
    

    View Slide

50. discovery 1
    big postgres is faster
    

    View Slide

51. discovery 2
    redis caching is slow
    

    View Slide

52. discovery 2
    !!!
    !!!
    

    View Slide

53. discovery 2
    totally unused, sadly
    

    View Slide

54. discovery 3
    threading is great!
    

    View Slide

55. threading is great!
    ...but deadlocks aren’t
    discovery 3
    

    View Slide

56. discovery 4
    usage is tricky
    

    View Slide

57. discovery 4
    metrics are tricky
    

    View Slide

58. discovery 4
    capacity is tricky
    

    View Slide

59. discovery 4
    monitor all of them
    

    View Slide

60. response time
    is all lies
    discovery 5
    

    View Slide

61. discovery 5
    check it from
    outside
    

    View Slide

62. so it’s all good now,
    right?
    

    View Slide

63. well… mostly…
    

    View Slide

64. well… mostly…
    !!!
    

    View Slide

65. time for the weird
    bug reports
    

    View Slide

66. is really, really hard
    synchronizing
    

    View Slide

67. is really, really slow
    background sync
    !!!
    

    View Slide

68. are pretty unreliable
    webooks
    

    View Slide

69. is pretty decent
    both
    

    View Slide

70. bundling is pretty fast
    after all that
    !
    !
    !
    

    View Slide

71. app time is two orders
    of magnitude faster
    2s → 20ms
    

    View Slide

72. it’s okay,
    I guess
    

    View Slide

73. we can do
    way better
    

    View Slide

74. we have a
    plan
    

    View Slide

75. ruby central
    grant
    

    View Slide

76. you
    can help
    

    View Slide

77. operators
    are standing by
    

    View Slide

78. thank you
    

    View Slide

79. questions?
    

    View Slide