Package io.nflow.engine.workflow.curated

Class BulkWorkflow

java.lang.Object
io.nflow.engine.model.ModelObject
io.nflow.engine.workflow.definition.WorkflowDefinition
io.nflow.engine.workflow.curated.BulkWorkflow
@Component public class BulkWorkflow extends WorkflowDefinition
Bulk child workflow executor that does not overflow the system.

  • Field Details

    • BULK_WORKFLOW_TYPE

      public static final String BULK_WORKFLOW_TYPE
      The type of default bulk workflow.
      See Also:

    • VAR_CHILD_DATA

      public static final String VAR_CHILD_DATA
      State variable name for child data.
      See Also:

    • VAR_CONCURRENCY

      public static final String VAR_CONCURRENCY
      State variable to define the maximum concurrency for executing child workflows.
      See Also:

    • SPLIT_WORK

      public static final WorkflowState SPLIT_WORK
      Bulk workflow states.

    • WAIT_FOR_CHILDREN_TO_FINISH

      public static final WorkflowState WAIT_FOR_CHILDREN_TO_FINISH

    • DONE

      public static final WorkflowState DONE

    • ERROR

      public static final WorkflowState ERROR

  • Constructor Details

    • BulkWorkflow

      protected BulkWorkflow(String type)
      Extend bulk workflow definition.
      Parameters:
      type - The type of the workflow.

    • BulkWorkflow

      public BulkWorkflow()
      Create bulk workflow definition.

  • Method Details

    • splitWork

      public NextAction splitWork(StateExecution execution, com.fasterxml.jackson.databind.JsonNode data)
      Call splitWorkImpl(io.nflow.engine.workflow.definition.StateExecution,com.fasterxml.jackson.databind.JsonNode) to create new child workflows.
      Parameters:
      execution - State execution context.
      data - Child data.
      Returns:
      Action to start waiting for children to finish or action to wait for children to be created.

    • splitWorkImpl

      protected boolean splitWorkImpl(StateExecution execution, com.fasterxml.jackson.databind.JsonNode data)
      Override this to create child workflows or add the children before starting the parent.
      Parameters:
      execution - State execution context.
      data - Child data.
      Returns:
      True to start processing the children, false to wait for children to be created.
      Throws:
      RuntimeException - Thrown by default implementation if children are not created before the parent is started.

    • waitForChildrenUntil

      protected org.joda.time.DateTime waitForChildrenUntil()
      Override this to customize the time to wait for new children before waking up. Default is one hour.
      Returns:
      Time when parent should wake up to check if new children have been added.

    • waitForChildrenToFinish

      public NextAction waitForChildrenToFinish(StateExecution execution, int concurrency)
      Check if all child workflows have finished. Start new children if needed and allowed by concurrency limit.
      Parameters:
      execution - State execution context.
      concurrency - The maximum number of child workflows to start.
      Returns:
      Action to retry this state at the time returned by waitForChildrenToCompleteUntil().

    • isRunning

      protected boolean isRunning(WorkflowInstance instance)
      Override this to determine if the child workflow is running or not. The default implementation returns true if the instance state is finished or created.
      Parameters:
      instance - The child workflow instance to check.
      Returns:
      True if the child is running, false otherwise.

    • waitForChildrenToCompleteUntil

      protected org.joda.time.DateTime waitForChildrenToCompleteUntil()
      Override this to customize the time to wait for children to finish. Default is 15 minutes. This is a safety mechanism only, as when the children normally go to finished state, they will automatically wake up the parent workflow.
      Returns:
      The time when parent should wake up to check if the children are finished.