Jobs Page

A job is a logical unit of work to be accomplished within the system. The Jobs page provides processing and debugging information for legacy jobs and flow runs. Use this page to inspect processing details, investigate halted jobs or failed flow runs, retry work, and download table data.

To access the Jobs page, go to Administration > Jobs.

Overview of job runs with IDs, types, statuses, and timestamps in an administration panel.

Required permissions

The page and actions described in this article require the following permissions.

Permission	Allows you to	Assigned by default to
View Jobs	View the Jobs page. Both the Legacy Jobs and Flows views require this permission.	System Admin
Retry Halted Jobs	Retry halted jobs from the Legacy Jobs table or failed flow runs from the Flows table. Without View Jobs, this permission applies only to the Flow Runs page.	System Admin; Business Admin
Fail Halted Jobs	Fail halted jobs from the Legacy Jobs table. This permission should be reserved for System Admins.	System Admin

Choose what to inspect

You can inspect legacy jobs or flow runs from the Jobs page. Choose the view that matches the type of processing work you need to investigate.

Jobs type filter

View	Use it when	How to open it
Legacy Jobs	You need to inspect cron jobs or machine tasks from the Document Processing Graph, including halted jobs.	Go to Administration > Jobs, and select Legacy Jobs from the drop-down list above the table.
Flows	You need to inspect submission-processing tasks that run as flows, including failed flow runs.	Go to Administration > Jobs, and select Flows from the drop-down list above the table.

This article focuses on the Jobs page. For information about user-facing flow runs, see Flow Runs Page.

Filter and search the table

The filters above the table apply to both views. By default, the table shows the work created today.

Filter	What it does
View selector	Switches between the Legacy Jobs and Flows tables.
Time range	Limits the table to work created in the selected period. You can add a time of day to the range. The selected period cannot exceed one week.
Type	Limits the table to one or more job types or flow-run types.
Status	Limits the table to one or more statuses. Apply a Failed filter to find flow runs you want to retry, or a Halted filter to find jobs you want to retry.
Search	Searches by correlation ID. Enter a value and press Enter.

To clear every filter and return to the default time range, click Clear All

Filters determine the scope of retry actions
The retry actions described in this article apply only to the work shown by the current filter.
The Fail halted jobs action is the exception: it applies to every halted job, regardless of the filter.

Jobs table filters

Understand the Legacy Jobs table

Each row in the Legacy Jobs table represents a job. The table lists jobs and their statuses and provides additional debugging details. It includes cron jobs and machine tasks from the Document Processing Graph.

Completed housekeeping jobs are excluded from the table. You can sort the table only by the Created column.

Metric	What it tells you
ID	The job's identifier.
Created	When the job was created.
Status	The job's current status.
Type	The job type.
Task	The task performed by the job. If the job belongs to a submission, this column links to the submission instead.
Started	When the job started.
Queue Time	The job's recorded queue time, in seconds.
Waiting Time	The job's recorded waiting time, in seconds.
Wall Time	The job's recorded wall time, in seconds.
User Time	The job's recorded user time, in seconds.
System Time	The job's recorded system time, in seconds.
Payload	The job's payload.

List of completed legacy jobs with details like ID, status, and timestamps.

Legacy job statuses

The Legacy Jobs table describes the following statuses.

Status	Meaning
Pending	The job is queued and waiting for a worker.
Scheduled	The job is scheduled to run at a later time.
Running	The job is currently running.
Completed	The job finished successfully.
Halted	The job stopped unexpectedly. Jobs are likely to halt because of an unexpected system failure, such as running out of disk space.
Failed	The job did not complete. A job changes to Failed when you retry it, when you fail it, or when it is canceled. When you retry a halted job, the original job changes to Failed, and the system creates a new job.

Inspect the system before retrying halted jobs
Inspect the system resources and the container before retrying a halted job. For more information, see Application Monitoring.

Inspect job details

To inspect a single job, click the drop-down menu () at the start of its row. The job details include the following attributes:

Attribute	What it tells you
Created, Started, Heartbeat, Finished	The job's timestamps.
Type, Worker Type, Worker, Task	What the job does and what processed it.
Payload, State Description, Pre Jobs	JSON details. For more information, see Inspect JSON details.
Parent ID, Join Job ID, Pending Pre Jobs, Failed Job ID, Retry Count	The job's place in the job graph and its retry history.

Retry halted jobs shown by the current filter

The Legacy Jobs table allows you to inspect halted jobs and retry them to complete the submission. Retrying applies only to the halted jobs currently shown by the filter, so apply a Halted status filter first to confirm which jobs are affected.

Apply the filters that match the jobs you want to retry.
Click Actions.
Click Retry halted jobs in filter in the drop-down menu.
Confirm the number of jobs shown in the dialog, and then click Retry.

The original jobs change to Failed, and the system creates new jobs. Retrying runs in the background, and a progress notification becomes available in the notifications menu.

The action is unavailable when the current filter returns no halted jobs.

Fail halted jobs

Do not fail a job unless Support advises you to
Failing a job has a side effect of continuing the job graph. A job graph is the sequence of jobs the system executes while processing a submission. We strongly recommend against failing a job unless the Support team advises you to do so.
Fail halted jobs applies to every halted job in the system, not only to the jobs shown by the current filter. You cannot limit its scope.

Click Actions.
Click Fail halted jobs in the drop-down menu.
Confirm the number of jobs shown in the dialog, and then click Fail.

The jobs change to Failed. The action is unavailable when there are no halted jobs.

Cancel active jobs

You cannot cancel active jobs from the Legacy Jobs table. Cancel them from the Running or Queued section of the Trainer instead.

To access the Trainer, go to Administration > Trainer.
For more information about the Trainer, see What is the Trainer?

Handle jobs that halt repeatedly

If you retry a halted job a couple of times and it always ends with a Halted status, contact the Support team. The Support team will investigate whether the submission is irreversibly broken and advise you on how to proceed.

If the submission is broken and cannot be processed, delete it from the Submissions page. For more information, see Navigating the Submissions Page.

Understand the Flows table

As part of Flows, Hyperscience has started running submission-processing tasks as flows rather than jobs. Each row in the Flows table represents a flow run.

You can sort the table only by the Started column.

Metric	What it tells you
ID	The flow run's identifier. Click it to open the flow run in the Flow Run Viewer.
Type	The flow-run type.
Started	When the flow run started.
Status	The flow run's current status.
Version	The version of the flow used for the run.

To inspect a single flow run without leaving the table, click the ellipsis button at the end of its row. The details show each task's Created and Started timestamps, its Type and Worker Type, its Payload, and, when the task did not complete, a Reason for Incompletion.

List of completed and failed flow runs with their IDs and statuses displayed.

Flow run statuses

Status	Meaning
Running	The flow run is in progress.
Completed	The flow run finished successfully.
Failed	The flow run did not complete. You can retry failed flow runs.
Cancelled	The flow run was canceled before it finished.
Paused	The flow run is paused.

Retry failed flow runs shown by the current filter

Retrying applies only to the failed flow runs currently shown by the filter.

Apply a Failed status filter, along with any other filters that match the flow runs you want to retry.
Click Actions.
Click Retry failed flow runs in filter in the drop-down menu.
Confirm the number of flow runs shown in the dialog, and then click Retry.

The system retries the failed flow runs shown by the filter. The action is unavailable when the current filter returns no failed flow runs.

Inspect JSON details

The JSON viewer makes large JSON payloads faster and easier to navigate when you inspect jobs, flow runs, and submission output.

On the Jobs page, the viewer appears inside the job details. Click the drop-down menu () on a row to open the details, and then read the JSON attributes in place.

The viewer shows a single line-numbered text view. Line numbers stay stable as you collapse and expand blocks. You can use the viewer to:

Search within the payload. Click the search bar above the payload, or press /, and then enter a value. The bar reports the number of matches, and you can step through them with Enter and Shift+Enter. Press Escape to clear the search. Matching covers collapsed blocks as well, and jumping to a match expands only what is needed to show it.
Collapse and expand individual objects and arrays instead of scrolling through the entire payload. Collapse controls appear in the gutter for payloads longer than a few lines.
Open or download referenced Hyperscience objects. URLs, endpoint references, and the UUIDs of recognized object types render as links. A download icon marks a link that saves a file rather than opening a page.
Copy or download the payload. Click Copy to copy it to the clipboard, or Download to save it as a JSON file.

The viewer remains responsive with large payloads because it renders only the lines currently in view.

Download table data

To download data from the table, click the menu button above the table's upper-right corner (), and then click Download Jobs JSON.

If you have filtered the table, the JSON includes only the data that meets the filter criteria. The JSON also includes only jobs or flow runs, depending on the current table view. The Legacy Jobs view saves the file as jobs.jsonand the Flows view saves it as workflows.json.