Class Interrupt

Terminates a flow by sending an interrupt signal. The interrupt activity can target the current flow (self-interrupt) or any other flow by its job ID (remote interrupt). Interrupted jobs have their status set to a value less than -100,000,000, indicating abnormal termination.

YAML Configuration — Self-Interrupt

When no target is specified, the activity interrupts the current flow. The flow terminates immediately after this activity executes. Use conditional transitions to route to an interrupt only when needed.

app:
  id: myapp
  version: '1'
  graphs:
    - subscribes: validation.check
      expire: 120

      activities:
        t1:
          type: trigger

        validate:
          type: worker
          topic: validate.input

        cancel:
          type: interrupt
          reason: 'Validation failed'
          throw: true
          code: 410
          job:
            maps:
              cancelled_at: '{$self.output.metadata.ac}'

        proceed:
          type: hook

      transitions:
        t1:
          - to: validate
        validate:
          - to: cancel
            conditions:
              code: 422               # interrupt only on validation failure
          - to: proceed

YAML Configuration — Remote Interrupt

When target is specified, the activity interrupts another flow while the current flow continues to transition to adjacent activities.

app:
  id: myapp
  version: '1'
  graphs:
    - subscribes: parent.flow
      expire: 120

      activities:
        t1:
          type: trigger
          job:
            maps:
              childJobId: '{$self.output.data.childJobId}'

        stop_child:
          type: interrupt
          topic: child.flow.topic     # topic of the target flow
          target: '{t1.output.data.childJobId}'
          throw: false                # do not throw (silent cancellation)
          descend: true               # also interrupt descendant sub-flows
          job:
            maps:
              interrupted: true

        done:
          type: hook

      transitions:
        t1:
          - to: stop_child
        stop_child:
          - to: done

Configuration Properties

Property Type Default Description
target string (current job) Job ID to interrupt. Supports @pipe expressions.
topic string (current topic) Topic of the target flow
reason string 'Job Interrupted' Error message attached to the interruption
throw boolean true Whether to throw a JobInterrupted error
descend boolean false Whether to cascade to child/descendant flows
code number 410 Error code attached to the interruption
stack string — Optional stack trace

Execution Model

  • Self-interrupt (no target): Category C. Verifies entry, maps job data, sets status to -1, and fires the interrupt. No children.
  • Remote interrupt (with target): Category B. Fires the interrupt best-effort, then uses executeLeg1StepProtocol to transition to adjacent activities.

See

InterruptActivity for the TypeScript interface

Hierarchy (view full)

Constructors

constructor

Properties

adjacencyList adjacentIndex code config context data engine guidLedger hook leg logger metadata status

Methods

assertGenerationalId authorizeEntry bindActivityData bindActivityError bindActivityMetadata bindActivityMetadataPaths bindActivityState bindDimensionalAddress bindJobError bindJobMetadata bindJobMetadataPaths bindJobState executeLeg1StepProtocol executeStepProtocol filterAdjacent getJobStatus getState getTriggerConfig initDimensionalAddress initPolicies initSelf interrupt interruptAnother interruptSelf isInterruptingSelf isJobComplete jobWasInterrupted mapInputData mapJobData mapOutputData mapStatusThreshold process processEvent registerTimeout resolveAdjacentDad resolveDad resolveInterruptOptions resolveStatus resolveThresholdHit setLeg setState setStatus shouldEmit verifyEntry verifyLeg1Entry verifyReentry

Constructors

constructor

Properties

adjacencyList

adjacencyList: StreamData[]

adjacentIndex

adjacentIndex: number = 0

code

code: number = 200

config

config: InterruptActivity

context

context: JobState

data

data: ActivityData

engine

engine: EngineService

guidLedger

guidLedger: number = 0

hook

hook: ActivityData

leg

leg: ActivityLeg

logger

logger: ILogger

metadata

metadata: ActivityMetadata

status

status: StreamStatus = StreamStatus.SUCCESS

Methods

assertGenerationalId

  • assertGenerationalId(jobGID, msgGID?): void

  • if the job is created/deleted/created with the same key, the 'gid' ensures no stale messages (such as sleep delays) enter the workstream. Any message with a mismatched gid belongs to a prior job and can safely be ignored/dropped.

    Parameters

    • jobGID: string
    • OptionalmsgGID: string

    Returns void

authorizeEntry

bindActivityData

bindActivityError

bindActivityMetadata

bindActivityMetadataPaths

bindActivityState

bindDimensionalAddress

bindJobError

  • bindJobError(data): void

  • unhandled activity errors (activities that return an ERROR StreamMessage status and have no adjacent children to transition to) are bound to the job

    Parameters

    • data: Record<string, unknown>

    Returns void

bindJobMetadata

bindJobMetadataPaths

bindJobState

executeLeg1StepProtocol

  • executeLeg1StepProtocol(delta): Promise<boolean>

  • Executes the 3-step Leg1 protocol for Category B activities (Leg1-only with children, e.g., Hook passthrough, Signal, Interrupt-another). Uses the incoming Leg1 message GUID as the GUID ledger key.

    Step A: setState + notarizeLeg1Completion + step1 markers (transaction 1) Step B: publish children + step2 markers + setStatusAndCollateGuid (transaction 2) Step C: if edge → runJobCompletionTasks + step3 markers + finalize (transaction 3)

    Parameters

    • delta: number

    Returns Promise<boolean>

    true if this transition caused the job to complete

executeStepProtocol

  • executeStepProtocol(delta, shouldFinalize): Promise<boolean>

  • Executes the 3-step Leg2 protocol using GUID ledger for crash-safe resume. Each step bundles durable writes with its concluding digit update in a single transaction.

    Parameters

    • delta: number
    • shouldFinalize: boolean

    Returns Promise<boolean>

    true if this transition caused the job to complete

filterAdjacent

getJobStatus

getState

getTriggerConfig

initDimensionalAddress

initPolicies

initSelf

interrupt

interruptAnother

interruptSelf

isInterruptingSelf

isJobComplete

jobWasInterrupted

mapInputData

mapJobData

mapOutputData

mapStatusThreshold

process

processEvent

registerTimeout

resolveAdjacentDad

resolveDad

resolveInterruptOptions

resolveStatus

resolveThresholdHit

setLeg

setState

setStatus

shouldEmit

verifyEntry

verifyLeg1Entry

  • verifyLeg1Entry(): Promise<boolean>

  • Leg1 entry verification for Category B activities (Leg1-only with children). Returns true if this is a resume (Leg1 already completed on a prior attempt). On resume, loads the GUID ledger for step-level resume decisions.

    Returns Promise<boolean>

verifyReentry

  • verifyReentry(): Promise<number>

  • Upon entering leg 2 of a duplexed activity. Increments both the activity ledger (+1) and GUID ledger (+1). Stores the GUID ledger value for step-level resume decisions.

    Returns Promise<number>

Settings

Member Visibility

On This Page

YAML Configuration — Self-InterruptYAML Configuration — Remote InterruptConfiguration PropertiesExecution Model
Constructors
Properties
Methods