WHEN MICROSERVICES MEET EVENT SOURCINGVinicius Gomes

1

VINICIUS GOMES

Software developer at ThoughtWorks

vvgomes.com/blog

twitter.com/vvgomes

2

AGENDA

The Traditional Approach

Introduction to Event Sourcing

CQRS

Demo

Summary

3

THE TRADITIONAL APPROACH

4

EXAMPLE

Online Restaurant

5

ONLINE RESTAURANT

Data Model

6

ONLINE RESTAURANT

Services

7

Restaurant

ONLINE RESTAURANT

8

📱POST

GET

I want to open an order!

200 ✔

Restaurant

ONLINE RESTAURANT

9

📱POST

GET

I want to add an item to

my order!

200 ✔

THE TRADITIONAL APPROACH

• Simple to implement

• Technology agnostic

10

GET POST PUT

DELETE

CHALLENGES

11

CHALLENGES

Coupling

12

CHALLENGES

Coupling

Resilience

13

Restaurant

📱POST

GET

I want to add an item to my order!

503👎

CHALLENGES

Coupling

Resilience

Response time

14

CHALLENGES

Coupling

Resilience

Response time

User intent

15

POST /orders/8659a0d6/items/

Add item to order

CHALLENGES

Coupling

Resilience

Response time

User intent

Current / mutable state

16

{ "id": “6a208c41…”, "status": "OPEN", "items": [ { "id": “be596c8e…”, "quantity": 1 }, { "id": “5d09509c…”, "quantity": 2 }, { "id": “cc52d1b6…”, "quantity": 1 } ] }

EVENT SOURCING

17

EVENT SOURCING

18

DOMAIN EVENTS

19

{ "payload": { "orderId": “be596c8e…”, "itemId": “5d09509c…”, "quantity": 1 }, "payloadType": “com.restaurant.ItemAddedToOrderEvent”, "timestamp": “2017-03-25 08:48:51 -03:00”, "revision": “2”, "aggregateId": “6a208c41…”

}

Model the past

Immutable

Permanent

EVENT STORE

Append only

Sequential

Long-lived

20

AGGREGATES

21

state = reduce(handle, {}, events)

Reducer function

Reducing events into state

AGGREGATES

Example

22

{ }

AGGREGATES

Example

23

payload: { id: “42” }

{ "id": “42”, "status": "OPEN", "items": [ ] }

AGGREGATES

Example

24

{ "id": “42”, "status": "OPEN", "items": [

{ "id": “43”, “quantity": 1

} ] }

payload: { orderId: “42”, itemId: “43”, quantity: 1 }

payload: { orderId: “42”, itemId: “44”, quantity: 2 }

AGGREGATES

Example

25

{ "id": “42”, "status": "OPEN", "items": [

{ "id": “43”, “quantity": 1

}, { "id": “44”, “quantity": 2

} ] }

AGGREGATES

Example

26

{ "id": “42”, "status": "OPEN", "items": [

{ "id": “44”, “quantity": 2

} ] }

payload: { orderId: “42”, itemId: “43” }

AGGREGATES

Example

27

{ "id": “42”, "status": "PLACED", "items": [

{ "id": “44”, “quantity": 2

} ] }

payload: { orderId: “42” }

AGGREGATES

Example

28

{ "id": “42”, "status": "PLACED", "items": [

{ "id": “44”, “quantity": 2

} ] }

Order Aggregate

COMMANDS

29

Add item to order

{ "type": “com.restaurant.AddItemToOrderCommand”,

"payload": { "orderId": “be596c8e…”, "itemId": “5d09509c…”, "quantity": 1 }, "timestamp": “2017-03-25 08:48:51 -03:00”

}

COMMANDS

30

(state, command) -> [event]

Handling a command

COMMANDS

31

How do clients send commands?

📱

I want to add an item to my order!

?

COMMANDS

32

Reviewing REST

Resource Collection

host/api/orders

COMMANDS

33

Reviewing REST

Resource

host/api/orders/8659a0d6

COMMANDS

34

Reviewing REST

Nested Resource

host/api/orders/8659a0d6/items/5d1a8457

COMMANDS

35

Reviewing REST

HTTP MethodsGET POST PUT/PATCH DELETE

COMMANDS

36

Reviewing REST

Hypermedia Formats

COMMANDS

37

Commands can be resources

COMMANDS

Option #1

Generic “Commands” nested-resource

38

POST /orders/8659a0d6/commands

{ “type": “AddItemToOrderCommand”, "menuItem": “/menu/items/8d30d99“, “quantity": 2 }

COMMANDS

Option #2

Command URI

39

POST /orders/8659a0d6/items/commands/add

{ "menuItem": “/menu/items/8d30d99“, “quantity": 2 }

COMMANDS

Command discoverability

40

EVENT SOURCING

41

42

Queries?

QUERIES

Bad news 😔

The Event Store is not good for queries

43

Most applications

•A few writes

•A lot of reads

CQRS

44

CQRS

Command Query Responsibility Segregation

45

Write ✍

Read 📖

Service

CQRS

46

📱

Command

Query

Publish

Listen

Service A

CQRS

47

Service B

CQRS

48

Command model •Command definitions

•Event definitions

•The aggregate

•Aggregate repository

•Write only API

Query model •Event listeners

•Query entities

•Repositories

•Read only API

DEMO

49

DEMO

Online Restaurant

50

Service

DEMO

Online Restaurant

51

EXAMPLE: ONLINE RESTAURANT

52

Online Restaurant

Publishes Listen to

EXAMPLE: ONLINE RESTAURANT

53

Online Restaurant

Publishes Listen to

Publishes Listen to

TECHNOLOGIES

Java

Spring Boot

Axon Framework

Spring Data REST

RabbitMQ

54

https://github.com/vvgomes/event-driven-restaurant

SUMMARY

55

MICROSERVICES + EVENT SOURCING

Benefits •History based queries

•Audit log by design

• Immutability

•User intent

•Decoupling

•Resilience

56

Challenges •Complexity

•Snapshots

•Upcasting

•Race conditions

•Event contracts

•Eventual consistency

WHEN MICROSERVICES MEET EVENT SOURCINGVinicius Gomes

57