Jobs
A job is a logical unit of work to be accomplished within the system.
Jobs are likely halted due to unexpected system failure (e.g. out of disk space). It is recommended that you inspect system resources and the container before re-running halted jobs. For more information, see Application Monitoring.
With the introduction of flows, we expect the idea of “jobs” to be replaced with “flows” in future versions of Hyperscience. As part of that transition, we've added a drop-down list to the Jobs page that allows you to view either Legacy Jobs or Flows.
Legacy Jobs table
The Legacy Jobs table lists all jobs and their respective statuses and provides additional details for debugging. The table includes cron jobs and machine tasks from the Document Processing Graph. You can view the Legacy Jobs table by going to Administration > Jobs and clicking on Jobs in the drop-down list at the top of the page.
Various metrics are provided with each table item:
- Created
- Status
- Type
- Task
- Started
- Queue Time
- Waiting Time
- Wall Time
- User Time
- System Time
- Payload
The Legacy Jobs table allows you to inspect halted jobs and re-run them to complete the submission. You can choose whether to retry all halted jobs or the halted jobs that are currently shown in the filter.
To retry all halted jobs, follow the steps below:
- Click Actions.
- Click Retry all halted jobs in the drop-down menu.
To retry halted jobs that are currently shown in the filter, follow the steps below:
- Click Actions.
- Click Retry halted jobs in filter in the drop-down menu.
If you re-run a halted job, the status of the original job will change to "Failed," and a new job will be created.
Note that you will not be able to cancel active jobs from the table; you will have to cancel them in the “Running” or “Queued” section of the Trainer.
- To navigate to the Trainer, go to Administration > Trainer.
- For more information about the Trainer, see What is the Trainer?
If you re-run a halted job a couple of times, and the job always ends with a “Halted” status, contact our Support team. The Support team will investigate whether the submission is irreversibly broken, and you will be advised on how to proceed. In cases where the submissions are broken and can’t be processed, we recommend deleting the submissions from the Submissions page. For more information about the Submissions page, see Navigating the Submissions Page.
Failing a job has the unexpected 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 you are advised to do so by our Support team.
Flows table
As part of Flows, we've started running submission-processing tasks as "flows" rather than jobs.
To see information about flow processes, click on Flows in the drop-down list above the table. In the Flows view, you can see the following metrics for each table item:
- Type
- Started
- Status
- Version
Currently, you can only sort the tasks in the table by their start time.
You can view failed flows by applying a Failed status filter to the Flows table. You can choose whether to retry all failed flow or the failed flows that are currently shown in the filter.
To retry all failed flows, follow the steps below:
- Click Actions.
- Click Retry all failed flows in the drop-down menu.
To retry the failed flows that are currently shown in the filter, follow the steps below:
- Click Actions.
- Click Retry failed flows in filter in the drop-down menu.
Downloading data
To download data from the table, click the menu button above the table's top-right corner ( ), and then click Download Jobs JSON.
If you have filtered the table, the JSON will only include the data that meets the filter criteria. Likewise, the JSON will only include data about jobs or flows, depending on your current view of the table.