The jBPM code itself is using the Apache License v2.0.

Some other components we integrate with have their own license:

jBPM now uses git for its source code version control system. The sources of the jBPM project can be found here (including all releases starting from jBPM 5.0-CR1):

https://github.com/kiegroup/jbpm

The source of some of the other components we integrate with can be found here:

If you’re interested in building the source code, contributing, releasing, etc. make sure to read this README.

Signing to jboss.org will give you access to the JBoss wiki, forums and JIRA. Go to https://www.jboss.org/ and click "Register".

The only form you need to sign is the contributor agreement, which is fully automated via the web. As the image below says "This establishes the terms and conditions for your contributions and ensures that source code can be licensed appropriately"

https://cla.jboss.org/

To be able to interact with the core development team you will need to use JIRA, the issue tracker. This ensures that all requests are logged and allocated to a release schedule and all discussions captured in one place. Bug reports, bug fixes, feature requests and feature submissions should all go here. General questions should be undertaken at the mailing lists.

Minor code submissions, like format or documentation fixes do not need an associated JIRA issue created.

https://issues.jboss.org/browse/DROOLS

https://issues.jboss.org/browse/JBPM

https://issues.jboss.org/browse/GUVNOR

With the contributor agreement signed and your requests submitted to JIRA you should now be ready to code :) Create a GitHub account and fork any of the Drools, jBPM or Guvnor repositories. The fork will create a copy in your own GitHub space which you can work on at your own pace. If you make a mistake, don’t worry blow it away and fork again. Note each GitHub repository provides you the clone (checkout) URL, GitHub will provide you URLs specific to your fork.

https://github.com/kiegroup

When writing tests, try and keep them minimal and self contained. We prefer to keep the DRL fragments within the test, as it makes for quicker reviewing. If their are a large number of rules then using a String is not practical so then by all means place them in separate DRL files instead to be loaded from the classpath. If your tests need to use a model, please try to use those that already exist for other unit tests; such as Person, Cheese or Order. If no classes exist that have the fields you need, try and update fields of existing classes before adding a new class.

There are a vast number of tests to look over to get an idea, MiscTest is a good place to start.

https://github.com/kiegroup/drools/blob/master/drools-compiler/src/test/java/org/drools/integrationtests/MiscTest.java

When you commit, make sure you use the correct conventions. The commit must start with the JIRA issue id, such as JBRULES-220. This ensures the commits are cross referenced via JIRA, so we can see all commits for a given issue in the same place. After the id the title of the issue should come next. Then use a newline, indented with a dash, to provide additional information related to this commit. Use an additional new line and dash for each separate point you wish to make. You may add additional JIRA cross references to the same commit, if it’s appropriate. In general try to avoid combining unrelated issues in the same commit.

Don’t forget to rebase your local fork from the original master and then push your commits back to your fork.

With your code rebased from original master and pushed to your personal GitHub area, you can now submit your work as a pull request. If you look at the top of the page in GitHub for your work area their will be a "Pull Request" button. Selecting this will then provide a gui to automate the submission of your pull request.

The pull request then goes into a queue for everyone to see and comment on. Below you can see a typical pull request. The pull requests allow for discussions and it shows all associated commits and the diffs for each commit. The discussions typically involve code reviews which provide helpful suggestions for improvements, and allows for us to leave inline comments on specific parts of the code. Don’t be disheartened if we don’t merge straight away, it can often take several revisions before we accept a pull request. Luckily GitHub makes it very trivial to go back to your code, do some more commits and then update your pull request to your latest and greatest.

It can take time for us to get round to responding to pull requests, so please be patient. Submitted tests that come with a fix will generally be applied quite quickly, where as just tests will often way until we get time to also submit that with a fix. Don’t forget to rebase and resubmit your request from time to time, otherwise over time it will have merge conflicts and core developers will general ignore those.

The workbench web application is using the pre-installed other security domain for authenticating and authorizing users (as specified in the WEB-INF/jboss-web.xml inside the WARs).

The application server uses by default property files based realms - Please note that this configuration is intended only for demo purposes (users, roles and passwords are stored in simple property files on the filesystem).

Authentication is configured in the standalone.xml file as follows:

These are the default users:

Authentication can be customized by using any of the following options:

Navigate into the workbench web application and click on the menu Home → Admin and selecting Users.

Example for Linux platforms - run the following command and follow the script instructions:

By default the demo setup makes use of Hibernate auto DDL generation capabilities to build up the complete database schema, including all tables, sequences, etc. This might not always be welcomed (by your database administrator), and thus the installer provides DDL scripts for most popular databases.

DDL scripts are provided for both jBPM and Quartz schemas although Quartz schema DDL script is only required when the timer service should be configured with Quartz database job store. See the section on timers for additional details.

This can be used to initially create the database schema, but it can also serve as the basis for any\ optimization that needs to be applied - such as indexes, etc.

jBPM installer ant script performs most of the work automatically and usually does not require additional attention but in case it does, here is a list of available targets that might be needed to perform some of the steps manually.

To start exploring the project:

The authoring perspective contains the hiring.bpmn2 process and a set of forms for each human task. Click these assets to explore. Notice that different editors open for different types of assets.

To build the Project:

Build & Deploy creates a new JAR artifact that is deployed to the runtime environment as a new deployment unit.

After successfully building and deploying your project, you can verify its presence in the Deployments tab. Click Deploy ▸ Deployments to do so.

You can find all the deployed units in the Deployments tab. When you Build & Deploy a project from the Project Editor, it is deployed using the default configurations. That means using the Singleton Strategy, the default Kie Base and the default Kie Session.

If you want a more advanced deployment, undeploy and re-deploy your artifacts using their GAV and selecting non-default settings. Then, you will be able to set a different strategy, or use a non-default Kie Base or Kie Session.

Once your artifact that contains the process definition is deployed, the Process Definition will become available in Process Management ▸ Process Definitions.

To create new process instances:

Click Process Management ▸ Process Definitions.

Start your instance:

The Process Definitions tab contains all the available process definitions in the runtime environment. In order to add new process definitions, build and deploy a new project.

Most processes require additional information to create a new process instance. This is done through forms. For this project, fill in the name of the candidate that is to be interviewed.

When you click Submit , you create a new process instance. This creates the first task, that is available for the Human Resources team. To see the task, you need to logout and log in as a user with the appropriate role assigned, that is someone from the Human Resources.

When you start the process, you can interact with the human tasks. To do so, click Task ▸ Tasks.

Note that in order to see the tasks in the task list, you need to belong to specific user groups, for which the task is designed. For example, the HR Interview task is visible only for the members of the HR group, and the Tech Interview Task is visible only to the members of the IT group.

The ProcessRuntime interface defines all the session methods for interacting with processes, as shown below.

The session provides methods for registering and removing listeners. A ProcessEventListener can be used to listen to process-related events, like starting or completing a process, entering and leaving a node, etc. Below, the different methods of the ProcessEventListener class are shown. An event object provides access to related information, like the process instance and node instance linked to the event. You can use this API to register your own event listeners.

A note about before and after events: these events typically act like a stack, which means that any events that occur as a direct result of the previous event, will occur between the before and the after of that event. For example, if a subsequent node is triggered as result of leaving a node, the node triggered events will occur inbetween the beforeNodeLeftEvent and the afterNodeLeftEvent of the node that is left (as the triggering of the second node is a direct result of leaving the first node). Doing that allows us to derive cause relationships between events more easily. Similarly, all node triggered and node left events that are the direct result of starting a process will occur between the beforeProcessStarted and afterProcessStarted events. In general, if you just want to be notified when a particular event occurs, you should be looking at the before events only (as they occur immediately before the event actually occurs). When only looking at the after events, one might get the impression that the events are fired in the wrong order, but because the after events are triggered as a stack (after events will only fire when all events that were triggered as a result of this event have already fired). After events should only be used if you want to make sure that all processing related to this has ended (for example, when you want to be notified when starting of a particular process instance has ended.

Also note that not all nodes always generate node triggered and/or node left events. Depending on the type of node, some nodes might only generate node left events, others might only generate node triggered events. Catching intermediate events for example are not generating triggered events (they are only generating left events, as they are not really triggered by another node, rather activated from outside). Similarly, throwing intermediate events are not generating left events (they are only generating triggered events, as they are not really left, as they have no outgoing connection).

jBPM out-of-the-box provides a listener that can be used to create an audit log (either to the console or the a file on the file system). This audit log contains all the different events that occurred at runtime so it’s easy to figure out what happened. Note that these loggers should only be used for debugging purposes. The following logger implementations are supported by default:

The KieServices lets you add a KieRuntimeLogger to your session, as shown below. When creating a console logger, the knowledge session for which the logger needs to be created must be passed as an argument. The file logger also requires the name of the log file to be created, and the threaded file logger requires the interval (in milliseconds) after which the events should be saved. You should always close the logger at the end of your application.

The log file that is created by the file-based loggers contains an XML-based overview of all the events that occurred at runtime. It can be opened in Eclipse, using the Audit View in the Drools Eclipse plugin, where the events are visualized as a tree. Events that occur between the before and after event are shown as children of that event. The following screenshot shows a simple example, where a process is started, resulting in the activation of the Start node, an Action node and an End node, after which the process was completed.

A common requirement when working with processes is ability to assign a given process instance some sort of business identifier that can be later on referenced without knowing the actual (generated) id of the process instance. To provide such capabilities, jBPM allows to use CorrelationKey that is composed of CorrelationProperties. CorrelationKey can have either single property describing it (which is in most cases) but it can be represented as multi valued properties set.

Correlation capabilities are provided as part of interface

that exposes following methods:

Correlation is usually used with long running processes and thus require persistence to be enabled to be able to permanently store correlation information.

In the following text, we will refer to two types of "multi-threading": logical and technical. Technical multi-threading is what happens when multiple threads or processes are started on a computer, for example by a Java or C program. Logical multi-threading is what we see in a BPM process after the process reaches a parallel gateway, for example. From a functional standpoint, the original process will then split into two processes that are executed in a parallel fashion.

Of course, the jBPM engine supports logical multi-threading: for example, processes that include a parallel gateway. We’ve chosen to implement logical multi-threading using one thread: a jBPM process that includes logical multi-threading will only be executed in one technical thread. The main reason for doing this is that multiple (technical) threads need to be be able to communicate state information with each other if they are working on the same process. This requirement brings with it a number of complications. While it might seem that multi-threading would bring performance benefits with it, the extra logic needed to make sure the different threads work together well means that this is not guaranteed. There is also the extra overhead incurred because we need to avoid race conditions and deadlocks.

In general, the jBPM engine executes actions in serial. For example, when the engine encounters a script task in a process, it will synchronously execute that script and wait for it to complete before continuing execution. Similarly, if a process encounters a parallel gateway, it will sequentially trigger each of the outgoing branches, one after the other. This is possible since execution is almost always instantaneous, meaning that it is extremely fast and produces almost no overhead. As a result, the user will usually not even notice this. Similarly, action scripts in a process are also synchronously executed, and the engine will wait for them to finish before continuing the process. For example, doing a Thread.sleep(…​) as part of a script will not make the engine continue execution elsewhere but will block the engine thread during that period.

The same principle applies to service tasks. When a service task is reached in a process, the engine will also invoke the handler of this service synchronously. The engine will wait for the completeWorkItem(…​) method to return before continuing execution. It is important that your service handler executes your service asynchronously if its execution is not instantaneous.

An example of this would be a service task that invokes an external service. Since the delay in invoking this service remotely and waiting for the results might be too long, it might be a good idea to invoke this service asynchronously. This means that the handler will only invoke the service and will notify the engine later when the results are available. In the mean time, the process engine then continues execution of the process.

Human tasks are a typical example of a service that needs to be invoked asynchronously, as we don’t want the engine to wait until a human actor has responded to the request. The human task handler will only create a new task (on the task list of the assigned actor) when the human task node is triggered. The engine will then be able to continue execution on the rest of the process (if necessary) and the handler will notify the engine asynchronously when the user has completed the task.

RuntimeManager has been introduced to simplify and empower usage of knowledge API especially in context of processes. It provides configurable strategies that control actual runtime execution (how KieSessions are provided) and by default provides following:

Runtime Manager is primarily responsible for managing and delivering instances of RuntimeEngine to the caller. In turn, RuntimeEngine encapsulates two the most important elements of jBPM engine:

Both of these components are already configured to work with each other smoothly without additional configuration from end user. No more need to register human task handler or keeping track if it’s connected to the service or not.

RuntimeEngine interface provides the most important methods to get access to engine components:

RuntimeManager will ensure that regardless of the strategy it will provide same capabilities when it comes to initialization and configuration of the RuntimeEngine. That means

On the other hand, RuntimeManager maintains the engine disposal as well by providing dedicated methods to dispose RuntimeEngine when it’s no more needed to release any resources it might have acquired.

Singleton strategy - instructs RuntimeManager to maintain single instance of RuntimeEngine (and in turn single instance of KieSession and TaskService). Access to the RuntimeEngine is synchronized and by that thread safe although it comes with a performance penalty due to synchronization. This strategy is similar to what was available by default in jBPM version 5.x and it’s considered easiest strategy and recommended to start with.

It has following characteristics that are important to evaluate while considering it for given scenario:

Per request strategy - instructs RuntimeManager to provide new instance of RuntimeEngine for every request. As request RuntimeManager will consider one or more invocations within single transaction. It must return same instance of RuntimeEngine within single transaction to ensure correctness of state as otherwise operation done in one call would not be visible in the other. This is sort of "stateless" strategy that provides only request scope state and once request is completed RuntimeEngine will be permanently destroyed - KieSession information will be removed from the database in case persistence was used.

It has following characteristics:

Per process instance strategy - instructs RuntimeManager to maintain a strict relationship between KieSession and ProcessInstance. That means that KieSession will be available as long as the ProcessInstance that it belongs to is active. This strategy provides the most flexible approach to use advanced capabilities of the engine like rule evaluation in isolation (for given process instance only), maximum performance and reduction of potential bottlenecks intriduced by synchronization; and at the same time reduces number of KieSessions to the actual number of process instances rather than number of requests (in contrast to per request strategy).

It has following characteristics:

Regular usage scenario for RuntimeManager is:

The complexity of knowing when to create, dispose, register handlers, etc is taken away from the end user and moved to the runtime manager that knows when/how to perform such operations but still allows to have a fine grained control over this process by providing comprehensive configuration of the RuntimeEnvironment.

As the name suggest, its primary responsibility is to deploy (and undeploy) units. Deployment unit is kjar that brings in business assets (like processes, rules, forms, data model) for execution. Deployment services allow to query it to get hold of available deployment units and even their RuntimeManager instances.

So typical use case for this service is to provide dynamic behavior into your system so multiple kjars can be active at the same time and be executed simultaneously.

Complete DeploymentService interface is as follows:

Upon deployment, every process definition is scanned using definition service that parses the process and extracts valuable information out of it. These information can provide valuable input to the system to inform users about what is expected. Definition service provides information about:

So definition service can be seen as sort of supporting service that provides quite a few information about process definition that are extracted directly from BPMN2.

While it usually is used with combination of other services (like deployment service) it can be used standalone as well to get details about process definition that do not come from kjar. This can be achieved by using buildProcessDefinition method of definition service.

Process service is the one that usually is of the most interest. Once the deployment and definition service was already used to feed the system with something that can be executed. Process service provides access to execution environment that allows:

At the same time process service is a command executor so it allows to execute commands (essentially on ksession) to extend its capabilities.

Important to note is that process service is focused on runtime operations so use it whenever there is a need to alter (signal, change variables, etc) process instance and not for read operations like show available process instances by looping though given list and invoking getProcessInstance method. For that there is dedicated runtime data service that is described below.

An example on how to deploy and run process can be done as follows:

As you can see start process expects deploymentId as first argument. This is extremely powerful to enable service to easily work with various deployments, even with same processes but coming from different versions - kjar versions.

Runtime data service as name suggests, deals with all that refers to runtime information:

Use this service as main source of information whenever building list based UI - to show process definitions, process instances, tasks for given user, etc. This service was designed to be as efficient as possible and still provide all required information.

Some examples:

There are two important arguments that the runtime data service operations supports:

These provide capabilities for efficient management result set like pagination, sorting and ordering (QueryContext). Moreover additional filtering can be applied to task queries to provide more advanced capabilities when searching for user tasks.

User task service covers complete life cycle of individual task so it can be managed from start to end. It explicitly eliminates queries from it to provide scoped execution and moves all query operations into runtime data service. Besides lifecycle operations user task service allows:

On top of that user task service is a command executor as well that allows to execute custom task commands.

Complete example with start process and complete user task done by services:

QueryService provides advanced search capabilities that are based on Dashbuilder DataSets. The concept behind it is that users are given control over how to retrieve data from underlying data store. This includes complex joins with external tables such as JPA entities tables, custom systems data base tables etc.

QueryService is build around two parts:

DashBuilder DataSets provide support for multiple data sources (CSV, SQL, elastic search, etc) while jBPM - since its backend is RDBMS based - focuses on SQL based data sets. So jBPM QueryService is a subset of DashBuilder DataSets capabilities to allow efficient queries with simple API.

Terminology

While QueryDefinition and QueryParam is rather straight forward, QueryParamBuilder and QueryResultMapper is bit more advanced and require slightly more attention to make use of it in right way, and by that take advantage of their capabilities.

QueryResultMapper

Mapper as the name suggest, maps data taken out from data base (from data set) into object representation. Much like ORM providers such as hibernate maps tables to entities. Obviously there might be many object types that could be used for representing data set results so it’s almost impossible to provide them out of the box. Mappers are rather powerful and thus are pluggable, you can implement your own that will transform the result into whatever type you like. jBPM comes with following mappers out of the box:

Each QueryResultMapper is registered under given name to allow simple look up by name instead of referencing its class name - especially important when using EJB remote flavor of services where we want to reduce number of dependencies and thus not relying on implementation on client side. So to be able to reference QueryResultMapper by name, NamedQueryMapper should be used which is part of jbpm-services-api. That acts as delegate (lazy delegate) as it will look up the actual mapper when the query is actually performed.

QueryParamBuilder

QueryParamBuilder that provides more advanced way of building filters for our data sets. By default when using query method of QueryService that accepts zero or more QueryParam instances (as we have seen in above examples) all of these params will be joined with AND operator meaning all of them must match. But that’s not always the case so that’s why QueryParamBuilder has been introduced for users to build their on builders which will provide filters at the time the query is issued.

There is one QueryParamBuilder available out of the box and it is used to cover default QueryParams that are based on so called core functions. These core functions are SQL based conditions and includes following

QueryParamBuilder is simple interface that is invoked as long as its build method returns non null value before query is performed. So you can build up a complex filter options that could not be simply expressed by list of QueryParams. Here is basic implementation of QueryParamBuilder to give you a jump start to implement your own - note that it relies on DashBuilder Dataset API.

Once you have query param builder implemented you simply use its instance when performing query via QueryService

Typical usage scenario

First thing user needs to do is to define data set - view of the data you want to work with - so called QueryDefinition in services api.

This is the simplest possible query definition as it can be:

Once we have the sql query definition we can register it so it can be used later for actual queries.

From now on, this query definition can be used to perform actual queries (or data look ups to use terminology from data sets). Following is the basic one that collects data as is, without any filtering

Above query was very simple and used defaults from QueryContext - paging and sorting. So let’s take a look at one that changes the defaults of the paging and sorting

Now let’s take a look at how to do data filtering

With that end user is put in driver seat to define what data and how they should be fetched. Not being limited by JPA provider nor anything else. Moreover this promotes use of tailored queries for your environment as in most of the case there will be single data base used and thus specific features of that data base can be used to increase performance.

Further examples can be found here.

ProcessInstanceMigrationService provides administrative utility to move given process instance(s) from one deployment to another or one process definition to another. It’s main responsibility is to allow basic upgrade of process definition behind given process instance. That might include mapping of currently active nodes to other nodes in new definition.

Migration does not deal with process or task variables, they are not affected by migration. Essentially process instance migration means a change of underlying process definition process engine uses to move on with process instance.

Even though process instance migration is available it’s recommended to let active process instances finish and then start new instances with new version whenever possible. In case that approach can’t be used, migration of active process instance needs to be carefully planned before its execution as it might lead to unexpected issues.Most important to take into account is:

Answers to these question might save a lot of headache and production problems after migration. Best is to always stick with backward compatible processes - like extending process definition rather than removing nodes. Though that’s not always possible and in some cases there is a need to remove certain nodes from process definition. In that situation, migration needs to be instructed how to map nodes that were removed in new definition in case active process instance is at the moment in such a node.

Node mapping is given as a map of node ids (UniqueIds that are set in the definition) where key is the source node id (from process definition used by process instance) to target node id (in new process definition).

Again, process or task variables are not affected by process instance migration at the moment.

ProcessInstanceMigrationService comes with several flavors of migrate operation:

Migration can either be performed for single process instance or multiple process instances at the same time. Multiple process instances migration is a utility method on top of single instance, instead of calling it multiple times, users call it once and then service will take care of the migration of individual process instances.

Deployment Service provides convinient way to put business assets to an execution environment but there are cases that requires some additional management to make them available in right context.

Activation and Deactivation of deployments

Imagine situation where there are number of processes already running of given deployment and then new version of these processes comes into the runtime environment. With that administrator can decide that new instances of given process definition should be using new version only while already active instances should continue with the previous version.

To help with that deployment service has been equipped with following methods:

This feature allows smooth transition between project versions whitout need of process instance migration.

Deployment synchronization

Prior to jBPM 6.2, jbpm services did not have deployment store by default. When embedded in jbpm-console/kie-wb they utilized sistem.git VFS repository to preserve deployed units across server restarts. While that works fine, it comes with some drawbacks:

With version 6.2 jbpm services come with deployment synchronizer that stores available deployments into data base, including its deployment descriptor. At the same time it constantly monitors that table to keep it in sync with other installations that might be using same data source. This is especially important when running in cluster or when jbpm console runs next to custom application and both should be able to operate on the same artifacts.

By default synchronization must be configured (when runing as core services while it is automatically enabled for ejb and cdi extensions). To configure synchronization following needs to be configured:

With this, deployments will be synchronized every 3 seconds with initial delay of two seconds.

Invoking latest version of project’s processes

In case there is a need to always work with latest version of project’s process, services allow to interact with various operations using deployment id with latest keyword. Let’s go over an example to better understand the feature.

Initially deployed unit is org.jbpm:HR:1.0 which has the first version of an hiring process. After several weeks, new version is developed and deployed to the execution server - org.jbpm:HR.2.0 with version 2 of the hiring process.

To allow callers of the services to interact without being worried if they work with latest version, they can use following deployment id:

this will alwyas find out latest available version of project that is identified by:

version comparizon is based on Maven version numbers and relies on Maen based algorithm to find the latest one.

Here is a complete example with deployment of multiple versions and interacting always with the latest:

As illustrated this provides very powerful feature when interacting with frequently chaning environment that allows to always be up to date when it comes to use of process definitions.

Processes can be created by using one of the following three methods:

Represents a script that should be executed in this process. A Script Task should have one incoming connection and one outgoing connection. The associated action specifies what should be executed, the dialect used for coding the action (i.e., Java, JavaScript or MVEL), and the actual action code. This code can access any variables and globals. There is also a predefined variable kcontext that references the ProcessContext object (which can, for example, be used to access the current ProcessInstance or NodeInstance, and to get and set variables, or get access to the ksession using kcontext.getKieRuntime()). When a Script Task is reached in the process, it will execute the action and then continue with the next node. It contains the following properties:

Note that you can write any valid Java code inside a script node. This basically allows you to do anything inside such a script node. There are some caveats however:

Represents an (abstract) unit of work that should be executed in this process. All work that is executed outside the process engine should be represented (in a declarative way) using a Service Task. Different types of services are predefined, e.g., sending an email, logging a message, etc. Users can define domain-specific services or work items, using a unique name and by defining the parameters (input) and results (output) that are associated with this type of work. Check the chapter on domain-specific processes for a detailed explanation and illustrative examples of how to define and use work items in your processes. When a Service Task is reached in the process, the associated work is executed. A Service Task should have one incoming connection and one outgoing connection.

Processes can also involve tasks that need to be executed by human actors. A User Task represents an atomic task to be executed by a human actor. It should have one incoming connection and one outgoing connection. User Tasks can be used in combination with Swimlanes to assign multiple human tasks to similar actors. Refer to the chapter on human tasks for more details. A User Task is actually nothing more than a specific type of service node (of type "Human Task"). A User Task contains the following properties:

A user task should define the type of task that needs to be executed (using properties like TaskName, Comment, etc.) and who needs to perform it (using either actorId or groupId). Note that if there is data related to this specific process instance that the end user needs when performing the task, this data should be passed as the content of the task. The task for example does not have access to process variables. Check out the chapter on human tasks to get more detail on how to pass data between human tasks and the process instance.

Represents the invocation of another process from within this process. A sub-process node should have one incoming connection and one outgoing connection. When a Reusable Sub-Process node is reached in the process, the engine will start the process with the given id. It contains the following properties:

A Business Rule Task Represents a set of rules that need to be evaluated. The rules are evaluated when the node is reached. A Rule Task should have one incoming connection and one outgoing connection. Rules are defined in separate files using the Drools rule format. Rules can become part of a specific ruleflow group using the ruleflow-group attribute in the header of the rule.

When a Rule Task is reached in the process, the engine will start executing rules that are part of the corresponding ruleflow-group (if any). Execution will automatically continue to the next node if there are no more active rules in this ruleflow group. As a result, during the execution of a ruleflow group, new activations belonging to the currently active ruleflow group can be added to the Agenda due to changes made to the facts by the other rules. Note that the process will immediately continue with the next node if it encounters a ruleflow group where there are no active rules at that time.

If the ruleflow group was already active, the ruleflow group will remain active and execution will only continue if all active rules of the ruleflow group has been completed. It contains the following properties:

A Sub-Process is a node that can contain other nodes so that it acts as a node container. This allows not only the embedding of a part of the process within such a sub-process node, but also the definition of additional variables that are accessible for all nodes inside this container. A sub-process should have one incoming connection and one outgoing connection. It should also contain one start node that defines where to start (inside the Sub-Process) when you reach the sub-process. It should also contain one or more end events. Note that, if you use a terminating event node inside a sub-process, you are terminating just that sub-process. A sub-process ends when there are no more active nodes inside the sub-process. It contains the following properties:

A Multiple Instance sub-process is a special kind of sub-process that allows you to execute the contained process segment multiple times, once for each element in a collection. A multiple instance sub-process should have one incoming connection and one outgoing connection. It waits until the embedded process fragment is completed for each of the elements in the given collection before continuing. It contains the following properties:

The start of the process. A process should have exactly one start node (none start node which does not have event definitions), which cannot have incoming connections and should have one outgoing connection. Whenever a process is started, execution will start at this node and automatically continue to the first node linked to this start event, and so on. It contains the following properties:

Allows you to create branches in your process. A Diverging Gateway should have one incoming connection and two or more outgoing connections. There are three types of gateway nodes currently supported:

It contains the following properties:

Allows you to synchronize multiple branches. A Converging Gateway should have two or more incoming connections and one outgoing connection. There are three types of splits currently supported:

It contains the following properties:

While the flow chart focuses on specifying the control flow of the process, it is usually also necessary to look at the process from a data perspective. Throughout the execution of a process, data can be retrieved, stored, passed on and used.

For storing runtime data, during the execution of the process, process variables can be used. A variable is defined by a name and a data type. This could be a basic data type, such as boolean, int, or String, or any kind of Object subclass (it must implement Serializable interface). Variables can be defined inside a variable scope. The top-level scope is the variable scope of the process itself. Subscopes can be defined using a Sub-Process. Variables that are defined in a subscope are only accessible for nodes within that scope.

Whenever a variable is accessed, the process will search for the appropriate variable scope that defines the variable. Nesting of variable scopes is allowed. A node will always search for a variable in its parent container. If the variable cannot be found, it will look in that one’s parent container, and so on, until the process instance itself is reached. If the variable cannot be found, a read access yields null, and a write access produces an error message, with the process continuing its execution.

Variables can be used in various ways:

Finally, processes (and rules) all have access to globals, i.e. globally defined variables and data in the Knowledge Session. Globals are directly accessible in actions just like variables. Globals need to be defined as part of the process before they can be used. You can for example define globals by clicking the globals button when specifying an action script in the Eclipse action property editor. You can also set the value of a global from the outside using ksession.setGlobal(name, value) or from inside process scripts using kcontext.getKieRuntime().setGlobal(name,value);.

Action scripts can be used in different ways:

Actions have access to globals and the variables that are defined for the process and the predefined variable kcontext. This variable is of type ProcessContext and can be used for several tasks:

jBPM supports multiple dialects, like Java, JavaScript and MVEL. Java actions should be valid Java code, same for JavaScript. MVEL actions can use the business scripting language MVEL to express the action. MVEL accepts any valid Java code but additionally provides support for nested accesses of parameters (e.g., person.name instead of person.getName()), and many other scripting improvements. Thus, MVEL expressions are more convenient for the business user. For example, an action that prints out the name of the person in the "requester" variable of the process would look like this:

Constraints can be used in various locations in your processes, for example in a diverging gateway. jBPM supports two types of constraints:

Rule constraints do not have direct access to variables defined inside the process. It is however possible to refer to the current process instance inside a rule constraint, by adding the process instance to the Working Memory and matching for the process instance in your rule constraint. We have added special logic to make sure that a variable processInstance of type WorkflowProcessInstance will only match to the current process instance and not to other process instances in the Working Memory. Note that you are however responsible yourself to insert the process instance into the session and, possibly, to update it, for example, using Java code or an on-entry or on-exit or explicit action in your process. The following example of a rule constraint will search for a person with the same name as the value stored in the variable "name" of the process:

Timers wait for a predefined amount of time, before triggering, once or repeatedly. They can be used to trigger certain logic after a certain period, or to repeat some action at regular intervals.

This is a simple example of a basic process with a script task only:

You can see that we start by calling the static createProcess() method from the RuleFlowProcessFactory class. This method creates a new process with the given id and returns the RuleFlowProcessFactory that can be used to create the process. A typical process consists of three parts. The header part comprises global elements like the name of the process, imports, variables, etc. The nodes section contains all the different nodes that are part of the process. The connections section finally links these nodes to each other to create a flow chart.

In this example, the header contains the name and the version of the process and the package name. After that, you can start adding nodes to the current process. If you have auto-completion you can see that you have different methods to create each of the supported node types at your disposal.

When you start adding nodes to the process, in this example by calling the startNode(), actionNode() and endNode() methods, you can see that these methods return a specific NodeFactory, that allows you to set the properties of that node. Once you have finished configuring that specific node, the done() method returns you to the current RuleFlowProcessFactory so you can add more nodes, if necessary.

When you are finished adding nodes, you must connect them by creating connections between them. This can be done by calling the method connection, which will link previously created nodes.

Finally, you can validate the generated process by calling the validate() method and retrieve the created RuleFlowProcess object.

When unit testing your process, you test whether the process behaves as expected in specific use cases, for example test the output based on the existing input. To simplify unit testing, jBPM includes a helper class called JbpmJUnitBaseTestCase (in the jbpm-test module) that you can use to greatly simplify your JUnit testing, by offering:

For example, consider the following "hello world" process containing a start event, a script task and an end event. The following JUnit test will create a new session, start the process and then verify whether the process instance completed successfully and whether these three nodes have been executed.

JbpmJUnitBaseTestCase acts as base test case class that shall be used for jBPM related tests. It provides four usage areas:

JbpmJUnitBaseTestCase supports all three predefined RuntimeManager strategies as part of the unit testing. It’s enough to specify which strategy shall be used whenever creating runtime manager as part of single test:

Above is more complete example that uses PerProcessInstance runtime manager strategy and uses task service to deal with user tasks.

User tasks can be used in combination with swimlanes to assign multiple human tasks to the same actor. Whenever the first task in a swimlane is created, and that task has an actorId specified, that actorId will be assigned to (all other tasks of) that swimlane as well. Note that this would override the actorId of subsequent tasks in that swimlane (if specified), so only the actorId of the first human task in a swimlane will be taken into account, all others will then take the actorId as assigned in the first one.

Whenever a human task that is part of a swimlane is completed, the actorId of that swimlane is set to the actorId that executed that human task. This allows for example to assign a human task to a group of users, and to assign future tasks of that swimlame to the user that claimed the first task. This will also automatically change the assignment of tasks if at some point one of the tasks is reassigned to another user.

The permisions matrix below summarizes the actions that specific user roles are allowed to do. On the left side, possible operations are listed while user roles are listed across the top of the matrix.

The cells of the permissions matrix contain one of three possible characters, each of which indicate the user role permissions for that operation:

Furthermore, the following words or abbreviations in the table header refer to the following roles:

User roles are assigned to users by the definition of the task in the jBPM (BPMN2) process definition.

The following matrix describes the authorizations for all operations which modify a task:

The matrix below describes the authorizations used when retrieving task information. In short, it says that all users which have any role with regards to the specific task, are allowed to see the task. This applies to all operations that are used to retrieve any type of information about the task.

Task service supports task listeners to be invoked upon various life cycle events happening on given task instance. In majority of cases task event listeners are used to intercept certain operation to perform additional logic - like storing task information in separate tables for business activity monitoring needs.

Task event listeners are pluggable and users can provide their own implementation of org.kie.api.task.TaskLifeCycleEventListener interface. There are beforeTask* and afterTask* methods that are invoked upon given event occured on a task instance.

TaskEvent (org.kie.api.task.TaskEvent) is the only argument available to the listener that provides access to:

In many cases implementors of task event listener need to have access to task variables (either input or output or both) to perform required operations. It can be done as described above (using various services and content marshaller helper) though that in many cases leads to code duplication in multiple listeners thus an extended support was added in 6.5 to simply use TaskContext to obtain that information.

Method loadTaskVariables can be used to populate both input and output variables of a given task by simple and single method call. That method is "no op" in case task variables are already set on a task.

To improve performance task variables are automatically set when they are available - usually given by caller on task service:

Other than that loadTaskVariables should be used to populate task variables.

Below is the data base model used by task service with all tables and their relationship illustrated.

The "Subtasks" feature is an experimental feature in the task service. This feature allows one task to have sub-tasks in a parent-child relationship. The parent task can auto-complete depending on the state of its children (and the subtask strategy used).

You can use it by setting the parentId of a task, either when creating the task manually via the task service or otherwise by setting the ParentId parameter of the task definition in the BPMN2 process definition.

Whenever a process is started, a process instance is created, which represents the execution of the process in that specific context. For example, when executing a process that specifies how to process a sales order, one process instance is created for each sales request. The process instance represents the current execution state in that specific context, and contains all the information related to that process instance. Note that it only contains the (minimal) runtime state that is needed to continue the execution of that process instance at some later time, but it does not include information about the history of that process instance if that information is no longer needed in the process instance.

The runtime state of an executing process can be made persistent, for example, in a database. This allows to restore the state of execution of all running processes in case of unexpected failure, or to temporarily remove running instances from memory and restore them at some later time. jBPM allows you to plug in different persistence strategies. By default, if you do not configure the process engine otherwise, process instances are not made persistent.

If you configure the engine to use persistence, it will automatically store the runtime state into the database. You do not have to trigger persistence yourself, the engine will take care of this when persistence is enabled. Whenever you invoke the engine, it will make sure that any changes are stored at the end of that invocation, at so-called safe points. Whenever something goes wrong and you restore the engine from the database, you also should not reload the process instances and trigger them manually to resume execution, as process instances will automatically resume execution if they are triggered, like for example by a timer expiring, the completion of a task that was requested by that process instance, or a signal being sent to the process instance. The engine will automatically reload process instances on demand.

The runtime persistence data should in general be considered internal, meaning that you probably should not try to access these database tables directly and especially not try to modify these directly (as changing the runtime state of process instances without the engine knowing might have unexpected side-effects). In most cases where information about the current execution state of process instances is required, the use of a history log is mostly recommended (see below). In some cases, it might still be useful to for example query the internal database tables directly, but you should only do this if you know what you are doing.

The jbpm-audit module contains an event listener that stores process-related information in a database using JPA. The data model itself contains three entities, one for process instance information, one for node instance information, and one for (process) variable instance information.

The ProcessInstanceLog table contains the basic log information about a process instance.

The NodeInstanceLog table contains more information about which nodes were actually executed inside each process instance. Whenever a node instance is entered from one of its incoming connections or is exited through one of its outgoing connections, that information is stored in this table.

The VariableInstanceLog table contains information about changes in variable instances. The default is to only generate log entries when (after) a variable changes. It’s also possible to log entries before the variable (value) changes.

The AuditTaskImpl table contains information about tasks that can be used for queries.

The BAMTaskSummary table that collects information about tasks that is used by BAM engine to build charts and dashboards.

The TaskVariableImpl table contains information about task variable instances.

The TaskEvent table contains information about changes in task instances. Operations such as claim, start, stop etc are stored here to provide time line view of events that happened to given task.

To log process history information in a database like this, you need to register the logger on your session like this:

To specify the database where the information should be stored, modify the file persistence.xml file to include the audit log classes as well (ProcessInstanceLog, NodeInstanceLog and VariableInstanceLog), as shown below.

All this information can easily be queried and used in a lot of different use cases, ranging from creating a history log for one specific process instance to analyzing the performance of all instances of a specific process.

This audit log should only be considered a default implementation. We don’t know what information you need to store for analysis afterwards, and for performance reasons it is recommended to only store the relevant data. Depending on your use cases, you might define your own data model for storing the information you need, and use the process event listeners to extract that information.

Process events are stored in the database synchronously and within the same transaction as actual process instance execution. That obviously takes some time especially in highly loaded systems and might have some impact on the database when both history log and runtime data are kept in the same database. To provide an alternative option for storing process events, a JMS based logger has been provided. It can be configured to submit messages to JMS queue instead of directly persisting them in the database. It can be configured to be transactional as well to avoid issues with inconsistent data in case of process engine transaction is rolled back.

This is just one of possible ways to configure JMS audit logger, see javadocs for AuditLoggerFactory for more details.

Process and task variables are stored in autdit tables by default although there are stored in simplest possible way - by creating string representation of the variable - variable.toString(). In many cases this is enough as even for custom classes used as variables users can implement custom toString() method that produces expected "view" of the variable.

Though this might not cover all needs, especially when there is a need for efficient queries by variables (both task and process). Let’s take as an example a Person object that has following structure:

while at first look this seems to be sufficient as the toString() methods provide human readable format it does not make it easy to be searched by. As searching through strings like "Person [name="john", age="34"] to find people with age 34 would make data base query very inefficient.

To solve the problem variable audit has been based on VariableIndexers that are responsible for extracting relevant parts of the variable that will be stored in audit log.

By default (indexer that takes the toString()) will prodce single audit entry for single variable, so it’s one to one relationship. But that’s not the only option: indexers (as can be seen in the interface) returns list of objects that are the outcome of single variable indexation.

To make our person queries more efficient we could build custom indexer that would take Person instance and index it into separate audit entries one representing name and the other representing age.

That indexer will then be used to index Person class only and rest of variables will be indexed with default (toString()) indexer. Now when we want to find process instances or tasks that have person with age 34 we simple refer to it as

There is not even need to use like based queries so data base can optimize the query and make it efficient even with big set of data.

Building and registering custom indexers

Indexers are supported for both process and task variables. though they are supported by different interfaces as they do produce different type of objects representing audit view of the variable. Following are the interfaces to be implemented to build custom indexers:

Implementation is rather simple, just two methods to be implemented

Once the implementation is done, it should be packaged as jar file and following file needs to be included:

Indexers are discovered by ServiceLoader mechanism and thus the META-INF/services files need. All found indexers will be examined whenever process or task variable is about to be indexed.

Only the default (toString() based) indexer is not discovered but added explicitly as last indexer to allow custom ones to take the precedence over it.

Special consideration need to be taken when embedding jBPM inside an application that executes in Container Managed Transaction (CMT) mode, for instance EJB beans. This especially applies to application servers that does not allow accessing UserTransaction instance from JNDI when being part of container managed transaction, e.g. WebSphere Application Server. Since default implementation of transaction manager in jBPM is based on UserTransaction to get transaction status which is used to decide if transaction should be started or not, in environments that prevent accessing UserTrancation it won’t do its job. To secure proper execution in CMT environments a dedicated transaction manager implementation is provided:

This transaction manager expects that transaction is active and thus will always return ACTIVE when invoking getStatus method. Operations like begin, commit, rollback are no-op methods as transaction manager runs under managed transaction and can’t affect it.

To configure this transaction manager following must be done:

With following configuration jBPM should run properly in CMT environment.

You need to make sure the necessary dependencies are available in the classpath of your application if you want to user persistence. By default, persistence is based on the Java Persistence API (JPA) and can thus work with several persistence mechanisms. We are using Hibernate by default.

If you’re using the Eclipse IDE and the jBPM Eclipse plugin, you should make sure the necessary JARs are added to your jBPM runtime directory. You don’t really need to do anything (as the necessary dependencies should already be there) if you are using the jBPM runtime that is configured by default when using the jBPM installer, or if you downloaded and unzipped the jBPM runtime artifact (from the downloads) and pointed the jBPM plugin to that directory.

If you would like to manually add the necessary dependencies to your project, first of all, you need the JAR file jbpm-persistence-jpa.jar , as that contains code for saving the runtime state whenever necessary. Next, you also need various other dependencies, depending on the persistence solution and database you are using. For the default combination with Hibernate as the JPA persistence provider and using an H2 in-memory database and Bitronix for JTA-based transaction management, the following list of additional dependencies is needed:

You can use the JPAKnowledgeService to create your knowledge session. This is slightly more complex, but gives you full access to the underlying configurations. You can create a new knowledge session using JPAKnowledgeService based on a knowledge base, a knowledge session configuration (if necessary) and an environment. The environment needs to contain a reference to your Entity Manager Factory. For example:

You can also use the JPAKnowledgeService to recreate a session based on a specific session id:

Note that we only save the minimal state that is needed to continue execution of the process instance at some later point. This means, for example, that it does not contain information about already executed nodes if that information is no longer relevant, or that process instances that have been completed or aborted are removed from the database. If you want to search for history-related information, you should use the history log, as explained later.

You need to add a persistence configuration to your classpath to configure JPA to use Hibernate and the H2 database (or your own preference), called persistence.xml in the META-INF directory, as shown below. For more details on how to change this for your own configuration, we refer to the JPA and Hibernate documentation for more information.

This configuration file refers to a data source called "jdbc/jbpm-ds". If you run your application in an application server (like for example JBoss AS), these containers typically allow you to easily set up data sources using some configuration (like for example dropping a datasource configuration file in the deploy directory). Please refer to your application server documentation to know how to do this.

For example, if you’re deploying to JBoss Application Server v5.x, you can create a datasource by dropping a configuration file in the deploy directory, for example:

If you are however executing in a simple Java environment, you can use the JBPMHelper class to do this for you (see below for tests only) or the following code fragment could be used to set up a data source (where we are using the H2 in-memory database in combination with Bitronix in this case).

You need to configure the jBPM engine to use persistence, usually simply by using the appropriate constructor when creating your session. There are various ways to create a session (as we have tried to make this as easy as possible for you and have several utility classes for you, depending for example if you are trying to write a process JUnit test).

The easiest way to do this is to use the jbpm-test module that allows you to easily create and test your processes. The JBPMHelper class has a method to create a session, and uses a configuration file to configure this session, like whether you want to use persistence, the datasource to use, etc. The helper class will then do all the setup and configuration for you.

To configure persistence, create a jBPM.properties file and configure the following properties (note that the example below are the default properties, using an H2 in-memory database with persistence enabled, if you are fine with all of these properties, you don’t need to add new properties file, as it will then use these properties by default):

If you want to use persistence, you must make sure that the datasource (that you specified in the jBPM.properties file) is initialized correctly. This means that the database itself must be up and running, and the datasource should be registered using the correct name. If you would like to use an H2 in-memory database (which is usually very easy to do some testing), you can use the JBPMHelper class to start up this database, using:

To register the datasource (this is something you always need to do, even if you’re not using H2 as your database, check below for more options on how to configure your datasource), use:

Next, you can use the JBPMHelper class to create your session (after creating your knowledge base, which is identical to the case when you are not using persistence):

Once you have done that, you can just call methods on this ksession (like startProcess) and the engine will persist all runtime state in the created datasource.

You can also use the JBPMHelper class to recreate your session (by restoring its state from the database, by passing in the session id (that you can retrieve using ksession.getId())):

How to use the web-based Workbench

Use the war from the workbench distribution zip that corresponds to your application server. The differences between these war files are mainly superficial. For example, some JARs might be excluded if the application server already supplies them.

The workbench stores its data, by default in the directory $WORKING_DIRECTORY/.niogit, for example wildfly-10.0.0.Final/bin/.niogit, but it can be overridden with the system property-Dorg.uberfire.nio.git.dir.

Here’s a list of all system properties:

To change one of these system properties in a WildFly or JBoss EAP cluster:

If the Workbench already contains Projects the examples can be imported with the "Import Example" button.

Alternatively, to importing an example, a new empty project can be created from the Project Authoring.

Select "Quick setup" and give the Project a name and optional description.

After a Project has been created you need to define Types to be used by your rules.

Select "Data Object" from the "Create New Asset" menu.

Set the name and select a package for the new type.

Click "+ add field" button and set a field name and type and click on "Create" to create a field for the type.

Click "Save" to update the model.

Select "DRL file" (for example) from the "Create New Asset" menu.

Enter a file name for the new rule.

Enter a definition for the rule.

The definition process differs from asset type to asset type.

The full documentation has details about the different editors.

Once the rule has been defined it will need to be saved in the same way we saved the model.

Once rules have been defined within a project; the project can be built and deployed to the Workbench’s Maven Artifact Repository.

To build a project select the "Build & Deploy" from the Project Authoring.

Click "Build & Deploy" to build the project and deploy it to the Workbench’s Maven Artifact Repository.

When you select Build & Deploy the workbench will deploy to any repositories defined in the Dependency Management section of the pom in your workbench project. You can edit the pom.xml file associated with your workbench project under the Repository View of the project explorer. Details on dependency management in maven can be found here : http://maven.apache.org/guides/introduction/introduction-to-dependency-mechanism.html

If there are errors during the build process they will be reported in the "Messages" panel.

Now the project has been built and deployed; it can be referenced from your own projects as any other Maven Artifact.

The full documentation contains details about integrating projects with your own applications.

A workbench is structured with Organization Units, VFS repositories and projects:

Organization units are useful to model departments and divisions.

An organization unit can hold multiple repositories.

Repositories are the place where assets are stored and each repository is organized by projects and belongs to a single organization unit.

Repositories are in fact a Virtual File System based storage, that by default uses GIT as backend. Such setup allows workbench to work with multiple backends and, in the same time, take full advantage of backend specifics features like in GIT case versioning, branching and even external access.

A new repository can be created from scratch or cloned from an existing repository.

One of the biggest advantage of using GIT as backend is the ability to clone a repository from external and use your preferred tools to edit and build your assets.

The workbench authenticates its users against the application server’s authentication and authorization (JAAS).

On JBoss EAP and WildFly, add a user with the script $JBOSS_HOME/bin/add-user.sh (or .bat):

There is no need to restart the application server.

The Workbench uses the following roles:

Create a user with the role admin and log in with those credentials.

After successfully logging in, the account username is displayed at the top right. Click on it to review the roles of the current account.

After logging in, the home screen shows. The actual content of the home screen depends on the workbench variant (Drools, jBPM, …​).

The Workbench is comprised of different logical entities:

Move the mouse pointer over the panel splitter (a grey horizontal or vertical line in between panels).

The cursor will by changing indicate it is positioned correctly over the splitter. Press and hold the left mouse button and drag the splitter to the required position; then release the left mouse button.

Projects often need external artifacts in their classpath in order to build, for example a domain model JARs. The artifact repository holds those artifacts.

The Artifact Repository is a full blown Maven repository. It follows the semantics of a Maven remote repository: all snapshots are timestamped. But it is often stored on the local hard drive.

By default the artifact repository is stored under $WORKING_DIRECTORY/repositories/kie, but it can be overridden with the system property-Dorg.guvnor.m2repo.dir. There is only 1 Maven repository per installation.

The Artifact Repository screen shows a list of the artifacts in the Maven repository:

To add a new artifact to that Maven repository, either:

The Asset Editor is the principle component of the workbench User-Interface. It consists of two main views Editor and Overview.