Skip to Content
Flower 3.7.1 is released 🎉 See what's new
DocsGuidesRun Sub-Processes
Guides· 6 min read· Updated November 2026

Sub-Processes

A sub-process lets a step of your process run another process model.
The main process waits at this step. As soon as the sub-process is finished, the main process continues. From the point of view of the main process, the sub-process behaves like one single task.

đź“–

Reference: BPMN Sub-Processes explains embedded sub-processes, call activities, event sub-processes, and ad-hoc sub-processes in the BPMN standard.

The main process waits at the call activity until the sub-process has finished, then it moves on.

Use sub-processes to:

  • Reuse a process in several other processes, for example an approval or an onboarding.
  • Keep large models readable: Move a part of the process into its own model.
  • Let different teams own different parts: Each model can be maintained and published on its own.

Call Activity or Embedded Sub-Process?

BPMN has two ways to draw a sub-process. Flower runs both, but in different ways:

BPMN elementWhat it does in Flower
Call activityRuns another process model. Use it whenever you want to start a sub-process.
Embedded sub-processGroups steps inside the same model. Its tasks are work items of the same process instance, as if they were drawn in the main process. Expanded or collapsed makes no difference.

Every task whose work type is mapped to a model starts that model; Flower recommends the call activity for it, because the diagram then shows that another process is called. This guide is about the call activity. Everything that differs between the two, feature by feature, is in What Works Where.

How It Works

  1. A process instance reaches a call activity.
  2. Flower creates a work item for it, like for any other task. Because the work type of this work item is mapped to a model, Flower turns the work item into a process instance of this model, the sub-process instance.
  3. The sub-process runs. Its tasks are work items like in every other process.
  4. When the sub-process instance is resolved, the main process continues after the call activity.

The work item of the call activity and the sub-process instance are the same Jira work item. So you see exactly one work item at this step, and its status tells you whether the sub-process is still running. The description of the work item says Flower Subprocess and names the parent process and the parent process instance. It is shown in a purple panel, tasks are blue, models and process instances green.

Set Up a Sub-Process

  1. Publish the model that you want to call. Only published models can be started.
  2. Map a work type to this model. Open the Flower Settings, go to the Jira Work Type Mapping tab, and select a space, a work type, and the model. See Setup Flower.
  3. Model a call activity in the main model. Select the call activity, open the Jira section in the properties panel, and choose its space and work type. The lists only contain the values of the Jira work type mapping, and a link opens the Jira Work Type Mapping tab of the Flower Settings to change the mapping. A call activity needs its own space and work type: unlike for tasks, they are not inherited from the swimlane or the model.
đź’ˇ

The Flower Modeler checks this when you validate the model: a call activity without space and work type, or with a combination that is not in the Jira work type mapping, is an error. A model that calls itself, directly or through other models, is an error too.

⚠️

A work type that is mapped to a model starts this model every time a work item of this type is created in this space, also for normal tasks. Use a mapped work type for call activities only. The validation of the Flower Modeler warns you if an ordinary task uses a mapped work type.

Descriptions and Placeholders

In the description of the work items of a sub-process (the documentation of an activity), the placeholder {{flower.parent}} is the key of the parent process instance, so you can refer to the work item of the main process. See User Tasks. The default description of a work item in a sub-process says that it runs as a sub-process of this parent.

Pass Data to a Sub-Process

Flower does not pass parameters to a sub-process. The sub-process is a process instance with its own data: the fields of its work item.

  • Fields of the call activity’s work item are the data of the sub-process instance. If the work type has a form, the form is part of the sub-process instance.
  • In the expressions of a gateway in the sub-process, pi is the sub-process instance, and par is the calling (parent) process instance. Use par to read data of the main process, for example ${par.fields.summary == 'New Request'}. See Decision Handling in Gateways.
  • In the main process, cur after the call activity is the work item of the sub-process instance. You can check its status and resolution on the next sequence flow: check the values in the properties panel (group Condition, see Choose a Status or Resolution), or write an expression such as ${cur.fields.resolution.name == 'Done'}.

Forms and Mandatory Fields

Flower creates the work item of a call activity automatically, without a user filling in a screen. So the work type of a call activity should have no mandatory fields (the validator marks them red), see Mandatory Fields.

This applies to sub-processes started from a process. If a user starts the process directly, mandatory fields are fine and even wanted: Create the work item with the Jira Create button, or submit a form (also a Jira Service Management request), as described in Forms.

If the sub-process needs data that nobody can give when it starts, let the called model wait for the fields with a conditional start event instead.

Versions

A sub-process starts with the version of the model that is published when the call activity is reached. If you publish a new version later, running sub-processes keep the version they started with. The next call starts the new version.

Ending a Sub-Process

  • Regular end: When the sub-process reaches its end event, Flower resolves its work item, and the main process continues.
  • Terminate end event in the sub-process: Ends only this sub-process. The main process continues as usual. The same applies to a terminate end event inside an embedded sub-process.
  • Terminate end event in the main process: Ends the main process and every running sub-process with all its open work items, also nested sub-processes. See BPMN End Events.
  • Resolution: Flower sets the Jira resolution if the workflow does not, see Which Status Flower Sets.
  • Resolving the work item manually: If you resolve the work item of a sub-process by hand, the main process continues. Open work items inside the sub-process stay open. Close them yourself or let the sub-process finish first.

Limits and Special Cases

Timers, errors, and escalations on a call activity, a loop back into it, nested calls, and what happens when the work type is not mapped or the sub-process cannot be started are described in the reference BPMN Sub-Processes.

Find Sub-Processes with JQL

A sub-process instance is a process instance that belongs to a node of the main process:

GoalJQL
All sub-processes of a process instanceflowerType = "Process Instance" AND flowerInstance = YOUR-INSTANCE-KEY
The sub-process at one call activityflowerType = "Process Instance" AND flowerInstance = YOUR-INSTANCE-KEY AND flowerNode = YOUR-ACTIVITY-ID
All running sub-processesflowerType = "Process Instance" AND flowerNode IS NOT EMPTY AND resolution is EMPTY

The sub-process instance counts as a process instance of the called model. In the model report of the calling model, it also appears as an open work item at its call activity. More queries are in Search Flower Data with JQL.

What’s Next?

Last updated on