Observe the System: Two Sides of the Same Coin @ Test Master Meetup

Observability is one of the buzzwords of the day. We have the 3 pillars, a plethora of tools, great books. We moved our monolithic apps that lived in the datacenter to distributed microservices in the cloud, with the promised land of cost savings and an Agile organisation. However, are we delivering more business value?

Today, we send metrics to the systems managed by the Operations Team and at the same time to the Business Intelligence systems. However, we have the will for our teams to have autonomy and ownership of the systems which they put into production. How can we merge the best of both worlds, giving autonomy to our development teams, have a reliable system and increase the business value delivered?

Join João on his journey, using collaboration techniques from the DDD community to discover the Domain Events that generated business value and how it is related to the technical components of the system. During his talk, he will address the challenges of bringing the two sides of the same coin together, creating a healthy environment among all parties.

Transcript

1. Observe the System: Two Sides of the
   Same Coin
   Test Masters Series – Testing in Production
   

   View Slide

2. 2
   Strategic Software Delivery
   Consultant @ Xebia
   DDDer, EventStormer, Tech
   Enthusiast
   Food lover, you can find me in a
   beach reading a book
   Expat @ The Netherlands
   @joaoasrosa
   

   View Slide

3. 3
   © João Rosa
   @joaoasrosa
   

   View Slide

4. 4
   How did I end up here?
   @joaoasrosa
   

   View Slide

5. 5
   And you?
   @joaoasrosa
   

   View Slide

6. 6
   Photo by Saketh Garuda on Unsplash
   @joaoasrosa
   

   View Slide

7. 7
   Well, we measure…
   @joaoasrosa
   

   View Slide

8. 8
   Photo by Kevin Grieve on Unsplash
   @joaoasrosa
   

   View Slide

9. 9
   Myself, with a confused look:
   Ok, how do you know the
   impact of the product
   with the users?
   @joaoasrosa
   

   View Slide

10. @joaoasrosa
    

    View Slide

11. That awkward silence…
    @joaoasrosa
    

    View Slide

12. 12
    Anti-pattern #1:
    Technical measurement
    without the user in mind
    @joaoasrosa
    

    View Slide

13. 13
    We are a DevOps team…
    @joaoasrosa
    

    View Slide

14. 14
    Photo by Olesya Grichina on Unsplash
    @joaoasrosa
    

    View Slide

15. 15
    Anti-pattern #2:
    Monitoring as single responsibility
    (aka Monitoring-as-a-Job - credits to Mike Julian)
    @joaoasrosa
    

    View Slide

16. 16
    We monitoring everything…
    @joaoasrosa
    

    View Slide

17. 17
    Photo by Markus Spiske on Unsplash
    @joaoasrosa
    

    View Slide

18. 18
    Myself, trying to get to the bottom of it:
    What are the benefits of
    the yellow hammer,
    in your landscape?
    @joaoasrosa
    

    View Slide

19. The yellow hammer is the best!
    @joaoasrosa
    

    View Slide

20. The yellow hammer is the best!
    Or do you think that purple hammer is better
    because ?
    @joaoasrosa
    

    View Slide

21. 21
    Anti-pattern #3:
    One tool to rule them all!
    @joaoasrosa
    

    View Slide

22. 22
    "Silos" by warrick1 is licensed under CC BY-ND 2.0
    @joaoasrosa
    

    View Slide

23. 23
    Photo by Yung Chang on Unsplash
    @joaoasrosa
    

    View Slide

24. 24
    Photo by NeONBRAND on Unsplash
    @joaoasrosa
    

    View Slide

25. 25
    Back to the Observability topic…
    @joaoasrosa
    

    View Slide

26. 26
    Photo by Christoph Schmid on Unsplash
    @joaoasrosa
    

    View Slide

27. 27
    Logging
    Metrics
    Tracing
    @joaoasrosa
    

    View Slide

28. 28
    From Site Reliability Engineering Book - https://landing.google.com/sre/sre-book/chapters/part3/
    @joaoasrosa
    

    View Slide

29. 29
    Not the focus today…
    @joaoasrosa
    

    View Slide

30. 30
    @joaoasrosa
    And more not listed here!
    

    View Slide

31. 31
    But what I use to
    connect all the dots…
    @joaoasrosa
    

    View Slide

32. 32
    To deliver
    a reliable and usable
    product that satisfies
    the users needs
    @joaoasrosa
    

    View Slide

33. 33
    @joaoasrosa
    

    View Slide

34. 34
    @joaoasrosa
    

    View Slide

35. 35
    @joaoasrosa
    

    View Slide

36. 36
    @joaoasrosa
    

    View Slide

37. 37
    @joaoasrosa
    

    View Slide

38. 38
    @joaoasrosa
    Bounded Context
    - Language boundaries
    - Specific model for the Problem Space
    - Isolate concepts (models)
    - Communicate over Domain Events
    From more check: http://domainlanguage.com/ddd/reference/
    

    View Slide

39. 39
    @joaoasrosa
    From Using domain analysis to model microservices - https://docs.microsoft.com/en-us/azure/architecture/microservices/model/domain-analysis
    

    View Slide

40. 40
    @joaoasrosa
    

    View Slide

41. 41
    @joaoasrosa
    

    View Slide

42. 42
    @joaoasrosa
    

    View Slide

43. 43
    @joaoasrosa
    

    View Slide

44. 44
    @joaoasrosa
    Pattern #1:
    Measure Events on the borders
    of the Bounded Context
    

    View Slide

45. 45
    @joaoasrosa
    Pattern #1 implementation examples:
    Events in a queue
    Calls to an API endpoint
    Events generated by a batch process
    

    View Slide

46. 46
    @joaoasrosa
    

    View Slide

47. 47
    @joaoasrosa
    

    View Slide

48. 48
    @joaoasrosa
    

    View Slide

49. 49
    @joaoasrosa
    

    View Slide

50. 50
    @joaoasrosa
    Pattern #2:
    Measure Event Flows concerning
    the Business Process(es)
    

    View Slide

51. 51
    @joaoasrosa
    Pattern #2 implementation example:
    Events tracing
    - Using OpenTracing API - https://opentracing.io/
    - Crafted, CorrelationID
    

    View Slide

52. 52
    @joaoasrosa
    

    View Slide

53. 53
    @joaoasrosa
    Pattern #3:
    Measure Events within the
    Bounded Context
    

    View Slide

54. 54
    @joaoasrosa
    Pattern #3 implementation examples:
    Structured logging
    Metrics
    

    View Slide

55. 55
    @joaoasrosa
    Let’s have a quizzzzz!
    

    View Slide

56. 56
    @joaoasrosa
    1. Buying a ticket for a movie (via an mobile app)
    2. The app reserves my seat for 10 minutes
    3. I can only pay using iDeal
    4. If I don’t pay within the 10 minute window
    the seats are released for other potential
    customer
    Scenario
    

    View Slide

57. 57
    @joaoasrosa
    What if I pay within the 10 minute window (let’s
    say 9:49) but in the meantime the seat was
    released (lag time between the payment
    acknowledgment)?
    Question
    

    View Slide

58. 58
    My suggestion…
    @joaoasrosa
    

    View Slide

59. 59
    @joaoasrosa
    Pattern #4:
    Use Pivotal Events to validate
    the option of implementing a process
    

    View Slide

60. 60
    @joaoasrosa
    Pattern #4 implementation examples:
    Metrics
    

    View Slide

61. 61
    @joaoasrosa
    Pattern #4 implementation examples:
    Metrics
    However, we use it to drive
    business value
    

    View Slide

62. 62
    @joaoasrosa
    How about visualisations?
    

    View Slide

63. 63
    @joaoasrosa
    How about visualisations?
    aka, Dashboards
    

    View Slide

64. 64
    @joaoasrosa
    Pattern #5:
    Start your Dashboards from a
    Functional perspective
    Later move to the Technical ones (drill-down),
    which supports the Functional Dashboards
    

    View Slide

65. 65
    @joaoasrosa
    Pattern #5 implementation examples:
    

    View Slide

66. 66
    Photo by Camylla Battani on Unsplash
    @joaoasrosa
    

    View Slide