Jobs and job execution states - Amazon IoT Core
Services or capabilities described in Amazon Web Services documentation might vary by Region. To see the differences applicable to the China Regions, see Getting Started with Amazon Web Services in China (PDF).

Jobs and job execution states

The following sections describe the lifecycle of an Amazon IoT job and the lifecycle of a job execution.

Job states

The following diagram shows the different states of an Amazon IoT job.


                        
                        Image showing the different states of an Amazon IoT job.

A job that you create using Amazon IoT Jobs can be in one of the following states:

  • SCHEDULED

    During the initial job or job template creation using the Amazon IoT console, CreateJob API, or CreateJobTemplate API, you can select the optional scheduling configuration in the Amazon IoT console or the SchedulingConfig in the CreateJob API or CreateJobTemplate API. When you start a scheduled job containing a specific startTime, endTime, and endBehavoir, the job status updates to SCHEDULED. When the job reaches your selected startTime or the startTime of the next maintenance window (if you selected job rollout during a maintenance window), the status will update from SCHEDULED to IN_PROGRESS and begin rollout of the job document to all devices in the target group.

  • IN_PROGRESS

    When you create a job using the Amazon IoT console or the CreateJob API, the job status updates to IN_PROGRESS. During job creation, Amazon IoT Jobs starts rolling out job executions to the devices in your target group. After all the job executions have rolled out, Amazon IoT Jobs waits for devices to complete the remote action.

    For information about concurrency and limits that apply to in-progress jobs, see Job limits.

    Note

    When an IN_PROGRESS job reaches the end of the current maintenance window, the rollout of the job document will stop. The job will update to SCHEDULED until the startTime of the next maintenance window.

  • COMPLETED

    A continuous job is handled in one of the following ways:

    • For a continuous job without the optional scheduling configuration selected, it's always in progress and continues to run for any new devices that are added to the target group. It will never reach a status state of COMPLETED.

    • For a continuous job with the optional scheduling configuration selected, the following is true:

      • If an endTime was provided, a continuous job will reach COMPLETED status when endTime has passed and all job executions have reached a terminal status state.

      • If an endTime was not provided in the optional scheduling configuration, the continuous job will continue to perform the job document rollout.

    For a snapshot job, the job status changes to COMPLETED when all of its job executions enter a terminal state, such as SUCCEEDED, FAILED, TIMED_OUT, REMOVED, or CANCELED.

  • CANCELED

    When you cancel a job using the Amazon IoT console, the CancelJob API, or the Job abort configuration, the job status changes to CANCELED. During job cancellation, Amazon IoT Jobs starts canceling previously created job executions.

    For information about concurrency and limits that apply to jobs that are being canceled, see Job limits.

  • DELETION_IN_PROGRESS

    When you delete a job using the Amazon IoT console or the DeleteJob API, the job status changes to DELETION_IN_PROGRESS. During job deletion, Amazon IoT Jobs starts deleting previously created job executions. After all job executions have been deleted, the job disappears from your Amazon account.

Job execution states

The following table shows the different states of an Amazon IoT job execution and whether the state change is initiated by the device or by Amazon IoT Jobs.

Job execution states and source
Job execution state Initiated by device? Initiated by Amazon IoT Jobs? Terminal status? Can be retried?
QUEUED No Yes No Not applicable
IN_PROGRESS Yes No No Not applicable
SUCCEEDED Yes No Yes Not applicable
FAILED Yes No Yes Yes
TIMED_OUT No Yes Yes Yes
REJECTED Yes No Yes No
REMOVED No Yes Yes No
CANCELED No Yes Yes No

The following section describes more about the states of a job execution that's rolled out when you create a job with Amazon IoT Jobs.

  • QUEUED

    When Amazon IoT Jobs rolls out a job execution for a target device, the job execution status is set to QUEUED. The job execution remains in the QUEUED state until:

    • Your device receives the job execution and invokes the Jobs API operations and reports the status as IN_PROGRESS.

    • You cancel the job or job execution, or when the abort criteria that you specified is met, and the status changes to CANCELED.

    • Your device is removed from the target group and the status changes to REMOVED.

    
                                Image showing how a queued job execution changes state to
                                    IN_PROGRESS and how a job can get REJECTED if the device doesn't
                                    accept the job creation request.
  • IN_PROGRESS

    If your IoT device subscribes to the reserved Job topics $notify and $notify-next, and your device invokes either the StartNextPendingJobExecution API or the UpdateJobExecution API with a status of IN_PROGRESS, Amazon IoT Jobs will set the job execution status to IN_PROGRESS.

    The UpdateJobExecution API can be invoked multiple times with a status of IN_PROGRESS. You can specify additional details about the execution steps using the statusDetails object.

    Note

    If you create multiple jobs for each device, Amazon IoT Jobs and the MQTT protocol don't guarantee order of delivery.

  • SUCCEEDED

    When your device successfully completes the remote operation, the device must invoke the UpdateJobExecution API with a status of SUCCEEDED to indicate that the job execution succeeded. Amazon IoT Jobs then updates and returns the job execution status as SUCCEEDED.

    
                                Image showing how an in-progress job execution can fail and
                                    how to retry the execution.
  • FAILED

    When your device fails to complete the remote operation, the device must invoke the UpdateJobExecution API with a status of Failed to indicate that the job execution failed. Amazon IoT Jobs then updates and returns the job execution status as Failed. You can retry this job execution for the device using the Job execution retry configuration.

    
                                Image showing how an in-progress job execution can fail and
                                    how to retry the execution.
  • TIMED_OUT

    When your device fails to complete a job step when the status is IN_PROGRESS, or when it fails to complete the remote operation within the timeout duration of the in-progress timer, Amazon IoT Jobs sets the job execution status to TIMED_OUT. You also have a step timer for each job step of an in-progress job and applies only to the job execution. The in-progress timer duration is specified using the inProgressTimeoutInMinutes property of the Job execution timeout configuration. You can retry this job execution for the device using the Job execution retry configuration.

    
                                Image showing how an in-progress job execution can time out
                                    and how to retry the execution.
  • REJECTED

    When your device receives an invalid or incompatible request, the device must invoke the UpdateJobExecution API with a status of REJECTED. Amazon IoT Jobs then updates and returns the job execution status as REJECTED.

  • REMOVED

    When your device is no longer a valid target for the job execution, such as when it's detached from a dynamic thing group, Amazon IoT Jobs sets the job execution status to REMOVED. You can re-attach the thing to your target group and restart the job execution for the device.

  • CANCELED

    When you cancel a job or cancel a job execution using the console or the CancelJob or CancelJobExecution API, or when the abort criteria specified using the Job abort configuration is met, Amazon IoT Jobs cancels the job and sets the job execution status to CANCELED.