Skip to content

Refactoring a monolith into microservices with Spring Cloud as demo'ed at developer week Nuremberg 2016

License

Notifications You must be signed in to change notification settings

deichten/dwx-spring-cloud

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

6 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Microservices with Spring Cloud

Code samples as of the talk held at 22nd of June at developer week 2016 in Nuremberg by Paul Vassu and Daniel Eichten.

All samples should build if you have Java 8 locally installed and just type ./mvnw clean spring-boot:run (Linux, Mac) or mvnw.cmd clean spring-boot:run (Windows). The samples are structured after the steps demoed. Be aware that for the later samples you have to have a MySQL Database running locally as done in the demo using a docker container, as well as you might want to use a forked and own version of the config repo providing your local settings, e.g. your API key for OpenWeatherMap.

General Note The used API key for OpenWeatherMap is invalid. Please go to there, register yourself and request a key. Basic usage is free.

1 Monolith

The is the start of the journey. Everything is contained in one single app and serving statics as well as services using an in-mem DB and the connection to OpenWeatherMap.

This example will use an in-mem h2 database and therefore no local DB. But be aware that data will be lost after restart of the app.

2 Simple Split

As a first refactoring step the service offering real services as well as the server serving static files are being split-up and exposed though a single reverse proxy.

Persistance will only be an in-mem h2 database.

3 Introducing Registry

To eliminate the need to configure everything statically we do introduce a registry server that helps the various other services to register and discover each other. This example also will make a use of an non-in-mem DB but will allow to run multiple instaces of each service behind Zuul by telling the services to just use a random port that is free.

Attention In order to run this as well as the subsequent examples you'll have to have a MySQL database running. For demo purposes we used the official docker image of MySQL 5.7. If you want to switch to a different DB you may also do this by exchanging the SQL driver dependency in the pom.xml of the services part – all the examples are using Spring Data JPA (with Hibernate as provider) and should therefore run with any DB supported.

Also be aware that you should start the services in a specific order. To mitigate any potential issues you should start the registry always first. Also note that you may run multiple instances of the services part cause it is configured to pick a random port.

4 Further Split up

To further split up the services by purpose we split the single service app of step 3 into one service offering member CRUD ops as well as one service talking to OpenWeatherMap.

Attention As in step 3 you should run the registry first.

5 Production Ready Setup

In the final step the same setup as in 4 is used but everything supported by distributed tracing using Zipkin as well as Turbine Collector getting all hystrix streams of all components and reporting accordingly.

When opening the turbine collector be sure to add the default cluster identifier to the turbine stream. Otherwise only parts of your cluster will show up (usually the service named first – so gateway in this case). link to open

Attention In order to run this example you will have to start the config server always first and should start the registry as a second service when the config server successfully started and is showing as UP on http://localhost:8888/health.

About

Refactoring a monolith into microservices with Spring Cloud as demo'ed at developer week Nuremberg 2016

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published