You are viewing the documentation for Blueriq 16. Documentation for other versions is available in our documentation directory.

The Case Engine manages all operations on cases, case data and tasks. It communicates with the Blueriq Runtime through asynchronous messages and a REST API.

The Case Engine is essential to use the event driven DCM architecture and depends on multiple other components to function properly.

This section explains how to install the Case Engine, how to configure it and which other components are needed to be installed.

Installation

The Case Engine currently is a Runtime with specific configuration. It is installed by deploying the Runtime WAR to a supported application server, see Installing Runtime. Alternatively you can use the Blueriq DCM Development Installation which includes the case engine with configuration.

Dependencies

The Case Engine depends on the following components:

The bootstrap.properties file needs to have the following content to enable these profiles:

bootstrap.propeties
spring.profiles.active=native,\
    case-engine,\
    customerdata-client,\
    dcm-lists-publisher,\
    externaldatasources,\
    process-sql-store,\
    timeline-event-publisher-amqp,\
    trace-event-publisher-amqp,\
    audit

Configuration

Specific for the case engine there are multiple properties that need to be set.

These properties are divided into three files:

  • application.properties: common properties that concern security, logging and the connection to the customer data
  • application-case-engine.properties: properties for several queues, MongoDB and Quartz
  • application-externaldatasources.properties: properties that enable the Case Engine to communicate with the Process SQL Store

Create the files in the additional config location of the case engine and copy and paste the corresponding properties into each file.

When using the Blueriq DCM Development Installation this configuration is included

application.properties
### Users ###
blueriq.security.auth-providers.local01.type=in-memory
blueriq.security.auth-providers.local01.users.location=users.properties
blueriq.security.auth-providers-chain=local01

### Customerdata ###
blueriq.customerdata-client.url=http://localhost:30002/customerdata/api/v1/
blueriq.customerdata-client.username=blueriq
blueriq.customerdata-client.password=welcome

### Exports ###
blueriq.exports.description=Exports
blueriq.exports.prefix=export
blueriq.exports.folder=exports
blueriq.exports.enabled=true

### Security settings ###
blueriq.security.csrf-protection.enabled=false

### Blueriq logging ###
#logging.level.com.aquima=DEBUG
#logging.level.com.blueriq=DEBUG
logging.file.name=logs/case-engine.log

### Default Queue configuration, can be ommitted when overruled at specific configuration in application-case-engine.properties ###
blueriq.default.rabbitmq.host=localhost
blueriq.default.rabbitmq.port=30010
blueriq.default.rabbitmq.username=guest
blueriq.default.rabbitmq.password=guest
blueriq.default.rabbitmq.virtualHost=/
blueriq.default.rabbitmq.ssl.enabled=false
application-case-engine.properties
blueriq.case.engine.concurrency.concurrent-consumers=1
blueriq.case.engine.concurrency.max-concurrent-consumers=1

blueriq.case.engine.data.mongodb.host=localhost
blueriq.case.engine.data.mongodb.port=30012
blueriq.case.engine.data.mongodb.database=caseEngine

blueriq.locking.mongodb.host=localhost
blueriq.locking.mongodb.port=30012
blueriq.locking.mongodb.database=locks

spring.quartz.job-store-type=memory

spring.quartz.properties.org.quartz.threadPool.class=org.quartz.simpl.SimpleThreadPool
spring.quartz.properties.org.quartz.threadPool.threadCount=2

For more information on concurrent consumers see Configuring RabbitMQ.


application-external-datasources.properties
#### Datasources  H2 ###
blueriq.datasource.process-sql-store.url=jdbc:h2:./h2db/dcm;AUTO_SERVER=TRUE
blueriq.datasource.process-sql-store.username=sa
blueriq.datasource.process-sql-store.password=
blueriq.datasource.process-sql-store.driverClassName=org.h2.Driver
blueriq.hibernate.process-sql-store.hbm2ddl.auto=update
blueriq.hibernate.process-sql-store.dialect=org.hibernate.dialect.H2Dialect
blueriq.hibernate.process-sql-store.criteria.literal_handling_mode=BIND
blueriq.hibernate.process-sql-store.globally_quoted_identifiers=true
blueriq.hibernate.process-sql-store.globally_quoted_identifiers_skip_column_definitions=true  

hbm2ddl.auto

When using a production database, please use 'validate' instead of 'update' in combination with the supplied database scripts.

application-dcm-lists-publisher.properties
# queue configuration from blueriq.default.rabbitmq can be overruled
blueriq.dcm.lists-publisher.rabbitmq.exchangeName=dcmListsEvents

Authentication

The synchronous operations are protected with basic authentication. To specify the credentials required to log in to the Case Engine, create a user in user.properties with the case-engine role.

users.properties
#format: USERNAME=PASSWORD,PRIVILEGE1,PRIVILEGE2,...
#example: admin=welcome,agent,customer,underwriter

# User for Case Engine
caseengine={noop}caseengine,case-engine

If you are using a different authentication provider, like LDAP, you can create a user in that provider instead of users.properties, as long as it has the proper role.

Event logging

It is possible to enable logging for events sent to the Case Engine. A line is logged when the Case Engine starts processing the event and when it is finished.

Events list

  • AbortTaskService
  • AbortTaskWithoutUnlocking
  • AssignTask
  • CompleteTask
  • CreateCase
  • DemuxedThrowMessageEvent
  • GetCaseInfo
  • GetTaskInfo
  • RemoveCase
  • ScheduledEvent
  • StartTask
  • ThrowMessage

How to enable

By default, the events are logged on INFO level. To specifically enable logging for the events only, configure the following package on INFO level:

application(-case-engine).properties
logging.level.com.blueriq.dcm.caseengine.service=INFO

Example output

StartTask event:

Finishing a case

Whenever the process for a case finishes, the case will be closed. This means that the following persistent parts of a case will be removed:

  • The process and its tasks and data in the process-sql-store
  • The case and its manual tasks in the DCM Lists Service's MongoDB database
  • The case document in MongoDB
  • Any remaining entries in the scheduler database
  • The case aggregate and the case data aggregate

We are aware that it may not be desirable to remove the aggregates, for example if your process doesn't have steps for archival. Therefore, we introduced a property to control this behavior:

application-case-engine.properties
blueriq.case.engine.data.general.remove-aggregates-on-case-close = false

The default value for this property is true.