khanavijit / event-driven-spring-boot

Example Application to demo various flavours of handling domain events in Spring Boot

Event Driven Applications with Spring Boot

This projects tries to capture various options you have when dealing with Event Driven Spring Boot applications. The follwing Spring Technologies are being used:

• Spring Boot
• Spring Cloud Stream Rabbit
• Spring Data JPA

These examples contain various different ways to model and deal with events:

• Complete aggregates / entities in the events
• REST Resource URLs in events
• Partial parsing / handling of events in consumers
• Events as Atom Feeds

Prerequisites

• You need to have Docker installed

How to run and install the example

In the root directory you need to

1. Compile everything with ./mvnw package
2. Start everything up with docker-compose up --build

Running on Kubernetes

Mind the KubernetesSetup.md file in the kubernetes directory

URLs and Ports

Each of the modules is it's own Spring Boot Application which can be accessed as follows:

Name	Application / Enpoint Type	URL
Application Process	9000	http://localhost:9000
Credit Application	9001	http://localhost:9001/credit-application
Customer	9002	http://localhost:9002/customer and http://localhost:9002/customer/feed
Scoring	9003	No UI
CreditDecision	9004	http://localhost:9004/credit-decision and http://localhost:9004/credit-decision/feed

Messaging Infrastructure & Domain Events

Public Events

CreditApplicationNumberGeneratedEvent

Source: application-process

Persisted in source: no

Consumers:

• credit-application
• credit-decision

Topic: CreditApplicationNumberGeneratedTopic

CreditApplicationEnteredEvent

Source: credit-application

Persisted in source: yes in its own Table via JPA

Consumers:

• application-process
• credit-decision

Topic: CreditApplicationEnteredTopic

CustomerCreatedEvent

Source: customer

Persisted in source: no

Consumers:

• application-process
• credit-decision

Topic: CustomerCreatedTopic

ScoringPositiveEvent

Source: scoring

Persisted in source: no

Consumers:

• application-process
• credit-decision

Topic: ScoringPositiveTopic

ScoringNegativeEvent

Source: scoring

Persisted in source: no

Consumers:

• application-process
• credit-decision

Topic: ScoringNegativeTopic

ApplicationDeclinedEvent

Source: credit-decision

Persisted in source: not as an event

Consumers:

• application-process

Topic: ApplicationDeclinedTopic

Internal Events

Credit-Application

• CreditDetailsEnteredEvent
• FinancialSituationEnteredEvent

Both events are stored Source: credit-application Storage: Own Table via JPA

Feeds

Customer Feed

Url: http://localhost:9002/customer/feed

Contains URLs to Customer Resources

Credit Decision Feed

Url: http://localhost:9004/credit-decision/feed

Contains Application Numbers that have been confirmed

Event Types being used

This demo shows various types of event types: Events with all the data, Events with Resource Urls and "Events" as Feeds

Events with all the data

Especially the CreditApplicationEnteredEvent falls into this category: it contains all of the data for the credit application such as the financial situation and the details of the actual credit. By consuming this event you will not need additional roundtrips to upstream systems

Other events that fall into this category are:

• ApplicationNumberGeneratedEvent
• ScoringNegativeEvent
• ScoringPositiveEvent
• ApplicationDeclinedEvent

Idea of Bounded Context:

Please take a close look at how the CreditApplicationEnteredEvent is being reflected in the scoring application. Yes, we take in all the payload from the broker but the public model of the event has a clear focus on the scoring context's view on the data.

Events with a Resource URL

These Events do not contain a lot of information. They may contian something like a business process identifier such as the applicationNumber in this example but for the purpose of this demo I refrained from doing that. So the CustomerCreatedEvent only contians the URL to the Customer REST Resource from which interested contexts can obtain the payload from.

"Events" via Feeds

Althoug the usage of feeds is no plain and pure event driven processing style I think that they come in handy when you are dealing with situations like these:

• you have issues with your message broker and firewalls and these issues can't be resolved easily
• you need to have an event replay functionality in place that enables consumers to restore their replicated data

You can find "Events via Feeds" in the customer and the credit-decision (see Feeds) applications.