arun-gupta/microservices

Java EE and Microservices

Refactor monolith to microservices

This workspace will show how to refactor a Java EE monolith to a microservices-based application. This will be accomplished in multiple phases:

1. Phase 1: Functional decomposition of a WAR into multiple `WAR`s. Use hardcoded service URIs.

2. Phase 2: Use service registration and discovery

3. Phase 3: Continuous integration

4. Phase 4: Continuous deployment

5. Phase 5: Proactive monitoring and scale in/out

6. Phase 6: Change the implementation stack of a service

Run the application

1. Start ZooKeeper

   docker run -d -p 2181:2181 fabric8/zookeeper

2. Download and unzip WildFly 9.0.0.Final

3. Run WildFly

   ./bin/standalone.sh

4. Deploy the application

   mvn -f microservice/pom.xml install -DskipTests

NoOps

Operational concerns are addressed in the following documents:

1. Service Discovery

2. Continuous Integration

3. Continuous Deployment

4. Service Monitoring

microservices's Issues

Setup ZooKeeper path caches

Here is the pseudo code:

        PathChildrenCache pcache = new PathChildrenCache(curatorFramework, znode, true);
        pcache.start();

        List<ChildData> childdata = pcache.getCurrentData();

More details: http://curator.apache.org/curator-recipes/path-cache.html

Add Snoop to register and discover services

Create SnoopRegistry qualifier in services module.
Implement ServiceRegistry interface in org.javaee7.wildfly.samples.services.snoop package.

Snoop: https://github.com/ivargrimstad/snoop

Use ELK stack for distributed logging

Details: http://blog.arungupta.me/getting-started-elk-stack-wildfly/

Add etcd to register and discover services

Read #21 for more details.

Use ServiceDiscovery for populating the URLs in template.xhtml

Use ServiceDiscovery for populating the URLs in template.xhtml. Should it be a separate module as its used in everest-web as well?

Dockerize Order service

Deploy Order service using Docker container

Related to #36 and #37.

R&D QBit

I also trying to find some conceptions with implementation of microservices in Java EE world,
and found some lib, please take look, for more idea =)
http://www.slideshare.net/slideshow/embed_code/46377880
https://github.com/advantageous/qbit

Refactor any business logic from everest-web to individual components

everest-web should mostly be a dumb proxy pulling all the pages together. Currently it has too much business logic.

This would need #15 to complete first.

Add shipping functionality

• OrderBean publish a message to a JMS queue
• New ShippingBean consumes the message and starts the shipping process
• Updates the status in Order table (add a new column?)

Dockerize User service

Deploy User service using Docker container

Related to #37 and #38.

Adding service discovery using Jube

@jstrachan will update the issue when its ready to be used.

ZooKeeper configuration

ZooKeeper IP address and port are hardcoded in org.javaee7.wildfly.samples.services.zookeeper.ZooKeeper. This need to be done using a configuration file or some other means.

Resource library contract should be deployed at a central location

Resource Library Template should be deployed at a central location as opposed to packaged with each WAR file. This will be attempted with #14, or after it is fixed.

Add WildFly Swarm profile

Add a new profile to create fat jars using WildFly Swarm

Add support OpenShift

will be grateful to see how to configure microservice App git repository with OpenShift git repository and run them in WildFly AS 8 on OpenShift

Add Kubernetes to register and discover services

Read #21 for more details

Add @Fixed qualifier for fixed service registration/discovery

Fixed service registration/discovery should be moved to services module. This can be done by adding @Fixed qualifier, just like @ZooKeeper.

Consul is CP, not AP

You can see this explicitly in the docs as compared to the underlying Serf lib that is AP

It forms a quorum managed cluster and attempts consistent data access.

https://www.consul.io/intro/vs/serf.html

Implement service discovery using ZooKeeper

Currently the URIs of each service is hardcoded in the application. This need to be implemented in org.javaee7.wildfly.samples.everest.ServiceDiscoveryZooKeeper.

Add Distributed transaction example

need example how to organize transaction in micro services, like service A need successful respond from depended services n & service n need successful respond from service x.
http://stackoverflow.com/questions/24676364/transactions-in-microservices

Implement User, Catalog, Order service in Node.js

This will show that MSA is technology/stack agnostic. You can pick only one of services as well.

Related to #34 and #35

Dockerize Catalog service

Deploy User service using Docker container

Related to #36 and #38.

Move the UI pages to each application

For the microservices-based application, all UI pages are in the everest module. These should be refactored to each individual component and then used instead.

Implement User, Catalog, Order service in Vert.x

This will show that MSA is technology/stack agnostic. You can pick only one of services as well.

Related to #33 and #35

Distributed logging using logentries

https://logentries.com/

Connectivity problems between zookeeper and other services in the dockerized version

Hi all,
I don't know if this is in fact a problem, because I might be doing something wrong.
What I have is:

• zookeeper running in its own docker container
• each service (order, catalog, user and everest) running in its own docker container
  With this in place I can not get any registry/discover of the services as the zookeeper service is not seen by other services.

Don't we miss the links tag in the pom file for the docker images to be built as in:

<image>
  ...
  <run>
	<ports>
		...
	</ports>
	<links>
		<link>zookeper-service:zoo</link>
	</links>
  </run>
</image>

Thank you for your support.
Kind Regards
Bruno Almeida

Implement User, Catalog, Order service in WildFly Swarm

This will show that MSA is technology/stack agnostic. You can pick only one of services as well.

Related to #33 and #34

@ZooKeeper CDI Extension

Move services module as a CDI extension

Fixed service discovery need to store the URIs in a shared persistent store

Fixed service discovery need to store the URIs in a shared persistent store. Otherwise each service will have its own ConcurrentHashMap.

Distrubuted transaction

Hi arun. Thanks a lot for that example. I have the same issue here. I think about to migrate our warehouse managemnt system ( based on JEE7 ) to microservice architecture for industry 4.0 approach. The problem I have is how to solve the distrubuted transactions. I'm sure you have some ideas ;-)

Add Consul to register and discover services

Read #21 for more details

Store user credentials in the application

• Store user credentials
• Allow login/logout
• Store address

Add different implementations of catalog, user, and order service

This will show how microservices enable polyglot architecture.

Consider MEAN, Scala/Akka, Ruby on Rails, Grails.

Setup ZooKeeper watches to check the status of the service

Here is the pseudo code:

        uris = curatorFramework.getChildren().usingWatcher(new Watcher() {
            @Override
            public void process(WatchedEvent we) {
                if (we.getType() == Event.EventType.NodeDeleted) {}
            }
        }).forPath(znode);

Automatic registration of REST endpoints with @ZooKeeper

@ZooKeeper qualifier can be specified over any REST endpoint. The endpoint will then register with ZooKeeper as a service.

Additionally, if @ZooKeeper is specified on an Application class then all endpoints in that application are automatically registered in ZooKeeper using a default name.

List of countries from REST API

Read the list of countries using REST API at: http://restcountries.eu/

Create foreign key constraint between Item and ItemType

Add docker-maven-plugin profile

Add a new profile to create Docker images for different microservices

Add Ribbon load balancer

https://github.com/Netflix/ribbon/wiki/Getting-Started

Distributed logging using Loggly

https://www.loggly.com/

REST endpoint URLs are hard coded

These need to be dynamically generated