Event Sourcing: Or, How I Learned to Stop Layering And Love the Event Log

Wondering what Event Sourcing is? How is this approach different from the way we normally write applications? Is adding a message bus enough to call what you're doing Event Sourcing? Over the past several years I have developed an app leveraging Event Sourcing for the Master Bowlers Association Nationals, and have learned some of the associated pros and cons of seeing ES in action. Using this app as a case study, we're going to answer some questions about ES; What challenges can it solve; What challenges can it cause; What did I do well... and where did I go wrong.

Transcript

1. Event Sourcing Or, How I Learned to Stop Layering And Love the Event Log
2. Chad Hurd @CaffGeek
3. Event Sourcing Or, How I Learned to Stop Layering And Love the Event Log Is your application in the gutter? Spare yourself the grief, and strike out using Event Sourcing. Come on....don't be a turkey -- Adam Rehill
4. Our Journey 1. The Problem 2. A Familiar Story 3. The Solution 4. Some Code 5. Wrapping Up
5. Part 1 The Problem
8. Data Coming In • 8 Provinces competing • 5 Team Divisions ◦ 5 people per team + coach • 5 Singles Divisions • 21 games • Over 4 days Totals 5040 individual game scores • Names • Averages • Room Assignments • Travel Information • Practice Schedule • Guests • Profiles • Lane Draws • Hotel • Bowling Centres • Sponsors • etc
9. Data Going Out
10. Data Going Out
11. Data Going Out
15. Part 2 A Familiar Story
22. The Layered Approach
25. Tradeoffs 1. Models are often used for both reading & writing 2. We lost the context of why the change happened 3. We can’t look into the past 4. Scaling our app
26. Why? domainModel.Name = dto.Name; domainModel.Age = dto.Age; domainModel.Gender = dto.Gender; domainModel.Address = dto.Address; domainModel.City = dto.City; domainModel.Province = dto.Province;
32. Sum Up Tradeoffs • Read vs Write • Capturing Why • The Past • Scaling • Read != Writing
33. Part 3 The Solution
35. Key Terms Commands Events Command Handlers Aggregates Read Models (Event Handlers)
38. What Do We Need To Build This Command Handlers Event Storage Event Handlers Read Model Storage Orchestration
39. The Write Side
40. Data Coming In
41. Resulting Events
42. The Read Side
43. Event Sourcing
45. ...what he actually said was “It can scarcely be denied that the supreme goal of all theory is to make the irreducible basic elements as simple and as few as possible without having to surrender the adequate representation of a single datum of experience.”
46. Part 4 Some Code
47. Entering a Score
48. We enter the scores, issuing a command
49. Processing the command
51. Resulting Events
52. Store the Events
53. The Event Store
54. Building our Read Models
56. Persisting the Read Model
59. Querying for data
61. The Read Model is Returned
63. Framework Message Dispatcher
67. Framework Message Dispatcher
68. What is my purpose? You pass the butter
70. Part 5 Wrapping Up
71. Event versioning Large aggregates Messaging can be complex once you scale Eventual Consistency can become a problem Disadvantages to Event Sourcing
72. Event Store is a historical account View state at any point in time Read is optimized to be fast Write is optimized to be correct Scaling is easier Advantages
73. What I Would Do Differently My aggregates don’t do much My aggregates are not correct CommandQueries are shared Testing is mostly absent
74. What Did I Get Right No Eventual Consistency No Message Bus Rebuilding Read Models Choosing Event Sourcing
75. What’s Next
76. Chad Hurd @CaffGeek Source: https://github.com/CaffGeek/mbacnationals