This example illustrates how you can build a complete onboarding solution by combining multiple services (based on business processes and decisions).
It consists of three independent services:
Note that both the hr and payroll services are implemented in the corresponding quarkus based example.
Users will typically only interact with the main onboarding service, but this one relies on the other two to manage some of the work.
This is the main onboarding service that exposes a single REST endpoint to onboard new employees.
The service is defined using a combination of 3 business processes to describe the steps required to onboard new employees. This includes one main process that serves as the overall entry point and two sub-processes to handle the interaction with HR and payroll.
You will need:
Please follow the instruction for each of the individual services. It is recommended to install them in given order.
mvn clean package spring-boot:run
Kogito runtime supports multiple persistence types, including Infinispan. In order to use the Infinispan based persistence, you need to have a Infinispan server installed and available over the network. The default configuration, expects the server to be running on:
infinispan.remote.server-list=localhost:11222
If you need to change it, you can do so by updating the application.properties file located in src/main/resources.
You can install Infinispan server by downloading version 12.x from the official website.
Once Inifispan is up and running you can build this project with -Ppersistence to enable additional processing during the build. Next you start it in exact same way as without persistence.
This extra profile in maven configuration adds additional dependencies needed to work with infinispan as persistent store.
Once the services are up and running, you can use the following example to interact with the service.
To start a new onboarding for the given employee, use following sample request:
curl -X POST -H 'Content-Type: application/json' -d '{"employee" : {"firstName" : "Mark", "lastName" : "Test", "personalId" : "xxx-yy-zzz", "birthDate" : "1995-12-10T14:50:12.123+02:00", "address" : {"country" : "US", "city" : "Boston", "street" : "any street 3", "zipCode" : "10001"}}}' http://localhost:8080/onboarding
This will onboard the given employee by the contacting the hr service (for validation, adding an employee identifier and manager, etc.) and the payroll service (for adding information about payroll, taxation and vacation days). It will return all employee information (including the newly generated information from the hr and payroll service) and the status of the onboarding request. Note that subsequent calls with the same information (employee personal ID) will result in a failed status as an employee can only be onboarded once.
Please note that the first execution of this request might take slightly longer, as some lazy loading and initialization of various components might still be required. Subsequent calls should be a lot faster though.
This project is configured to run with JKube Maven plugin. You can simply run:
# make sure that the docker env from minikube is enabled locally $ eval $(minikube -p minikube docker-env) # build the service, the image and deploy it on Minikube: $ mvn clean install k8s:deploy -Pkubernetes
If you wish to only generate the image, you can use k8s:build instead. To generate the Kubernetes resources use k8s:resource. See the official JKube documentation for more information.
NOTE: If you're targeting a Kubernetes or OpenShift cluster, use the resources created on
target/classes/META-INF/jkubedirectory.
You can use the default configuration to deploy on Minikube and start exploring the examples right way. If you're planning to deploy on an actual cluster, review the src/main/jkube files to make sure the permissions added to the service are fine.
Unfortunately, JKube does not support Knative natively. What you can do is use this example as a reference:
apiVersion: serving.knative.dev/v1 kind: Service metadata: name: onboarding-springboot spec: template: spec: containers: - image: kogito/onboarding-springboot:latest imagePullPolicy: IfNotPresent serviceAccountName: onboarding-springboot ports: - containerPort: 8080 name: http protocol: TCP
Run the commands bellow to deploy the Knative service:
# make sure that the docker env from minikube is enabled locally $ eval $(minikube -p minikube docker-env) # build the service, the image and the other resources: $ mvn k8s:resource -Pkubernetes # apply the other resources $ kubectl apply -f target/classes/META-INF/jkube/onboarding-springboot-service.yml $ kubectl apply -f target/classes/META-INF/jkube/onboarding-springboot-serviceaccount.yml $ kubectl apply -f target/classes/META-INF/jkube/onboarding-springboot-view-rolebinding.yml # apply the knative service resource $ kubectl apply -f <path-to-file>.yaml
NOTE: Please install the Knative Serving platform before attempting to run the command above.
In addition, following permissions should be added to the default service account so the service discovery can properly happen:
oc policy add-role-to-group view system:serviceaccounts:onboarding-service -n istio-system oc policy add-role-to-group knative-serving-core system:serviceaccounts:onboarding-service -n default
If using Knative Serving Operator on OpenShift 4.x, the permissions that need to be set are
oc adm policy add-role-to-user knative-serving-core system:serviceaccount:myproject:onboarding-service -n default oc adm policy add-role-to-user view system:serviceaccount:myproject:onboarding-service -n istio-system
Once you have deployed the three services, you can run the following command to expose the onboarding-springboot via Node Port (the default option is already configured for you):
minikube service --url onboarding-springboot
You should see the URL in the terminal. Use it to make the calls to the onboarding-springboot