Skip to content

Latest commit

 

History

History
184 lines (135 loc) · 8.95 KB

README.md

File metadata and controls

184 lines (135 loc) · 8.95 KB
Raw

Namespaces for Zeebe (Camunda 8) Applications

If you apply process orchestration with Camunda over a huge company, it is hard to avoid name clashes of task types and called processes. Several teams may use the same name for different task types and/or worker implementations or in call activities.

Namespaces for each project helps you here.

This project offer some configuration mechanisms to set the namespace in a single place and apply it to all process models and workers used in this spring boot project.

Adding pre- and postfix to process ID

As there is no hook in the Zeebe client to change the BPMN diagram before deploying, you can use the Maven-Resources-Plugin to adjust the process model with a namespace.

The process ID has to be a valid XML ID, so the characters "@" and "${}" are not allowed here.

But you can define another delimiter like "AAA" to change the process ID with Maven before is gets packed into the war files.

The process ID contains the value AAAprocess.tagAAA_deliveryProcess. The Maven resources plugin evaluates it to processOne_1_0_deliveryProcess. The following configuration chain is required to change the process ID.

Maven-Resources-Plugin configuration

<plugin>
  <artifactId>maven-resources-plugin</artifactId>
  <configuration>
    <delimiters>
      <delimiter>${*}</delimiter>
      <delimiter>@</delimiter>
      <delimiter>AAA</delimiter>
    </delimiters>
  </configuration>
</plugin>

Don't forget to enable filtering to replace the variables with their values.

pom.xml properties

<properties>
  <process.key>processOne</process.key>
  <process.version>1_0</process.version>
  <process.tag>${process.key}_${process.version}</process.tag>
</properties>

Adding pre- and postfix to task types

Without any configuration, as task type _singleWork is picked up by a worker with the annotation @JobWorker(type = "_myTaskType").

The connection between task type and worker annotation can be configured using an ID tag. The value will be put in front or appended to the worker type. And the variables for ID tag value are injected when a new process instance is started.

As an example put the following values into the application.properties configuration

zeebe.namespace.mappings.processKey=processOne
zeebe.namespace.mappings.version=1_0

Now the worker from above registers to tasks with type processOne_1_0_myTaskType.

To create a task with the type, the type field in the modeler requires a FEEL expression: = pt + "_singleWork". pt is the default value for the zeebe.namespace.conf.tagName configuration.

The value for the variable pt is set automatically by the starter.

Adding pre- and postfix to call activities

To create the called process, you have to apply the same tasks as for the processes.

To call the processes from the super processes, add a FEEL expression for the Process ID of the called element like pt + "subProcessOne.

Given the same value for pt as above, the expression gets evaluated to processOne_1_0_subProcessOne.

Complete configuration properties

ID tag

The ID tag gets calculated connecting the variables listed in zeebe.namespace.variableNames and zeebe.namespace.typeDelimiter, configured by the zeebe.namespace.tagValueDelimiterMode. It takes the values for the variables from zeebe.namespace.mappings.<key>.

tagValueDelimiterMode Explanation value for pt
AS_PREFIX Adds the delimiter to the end of the ID tag. The task type can be given without a FEEL expression processOne_1_0_
AS_POSTFIX Adds the delimiter in front of the ID tag. The task type can be given without a FEEL expression _processOne_1_0
DEPENDS_TYPE_MODE If typeMode is set to AUTO, EXPLICIT or OFF, no delimiter will be added to the ID tag. If typeMode is set to PREFIX or POSTFIX, see AS_PREFIX or AS_POSTFIX See other values
UNTOUCHED The delimiter will not be added processOne_1_0

Type mode

The zeebe.namespace.typeMode configuration influences how the @JobWorker type annotation is processed and registered.

These values are defined:

  • AUTO: it checks if the type value begins and/or ends with the value of the typeDelimiter. At each position, the value of the ID tag will be inserted. Examples:

    type value result
    _singleWork processOne_1_0_singleWork
    singleWork_ singleWork_processOne_1_0
    _singleOne_ processOne_1_0_singleWork_processOne_1_0

  • PREFIX: the ID tag will always be inserted in front of the type, connected with the value of the typeDelimiter. The delimiter must not be given in the process definition.

    type value result
    singleWork processOne_1_0_singleWork

  • POSTFIX: The ID tag will always be appended to the type, connected with the value of the typeDelimiter. The delimiter must not be given in the process definition.

    type value result
    singleWork singleWork_processOne_1_0

  • EXPLICIT: No ID tag will be calculated, only the referenced variables will be replaced. Given the configuration zeebe.namespace.mappings.test=done.

    type value result
    singleWork_${test} singleWork_done

  • OFF: The type values will be used without any replacement. It may help with debugging the setup, as all replacement will the logged on startup.

Configuration reference

Property Range of values Description Default value
zeebe.namespace.mappings.<key> String Creates a variable in the list with name <key> and value.
zeebe.namespace.tagName String Gets calculated from tagVariableNames and typeDelimiter to create the ID-tag pt
zeebe.namespace.tagValueDelimiterMode AS_PREFIX, AS_POSTFIX, DEPENDS_TYPE_MODE, UNTOUCHED Where should the value of tagDelimiter be placed DEPENDS_TYPE_MODE
zeebe.namespace.tagVariableNames String list Comma delimited list of variable names to create the ID-tag with typeDelimiter processKey, version
zeebe.namespace.typeDelimiter String The delimiter between variables _ (underscore)
zeebe.namespace.typeMode AUTO, PREFIX, POSTFIX, EXPLICIT, OFF See below AUTO
zeebe.namespace.variableMode ALL, FULLTAG, TAGONLY, OFF Defines to inject variables into started process instances. More details below. ALL
zeebe.namespace.variableStart String Begin of a variable reference in the type field ${
zeebe.namespace.variableEnd String End of a variables reference in the type field }

Examples

See the subfolder examples for executable examples.

Credentials

The code of this project comes from Bernd Köcke from 1&1.