Tracking the Progress of Compute Jobs
Data Consumers can check the progress of the compute job transactions via the Single Compute Jobs or Multiple Compute Jobs dashboard.
To check the status of a compute job:
- Log in to the marketplace.
- From the left navigation bar, go to My Orders > Single Compute jobs or Multiple Compute Jobs.
- For single compute jobs: View the progress of your compute job under the 'Compute Job Status' column. The status of the compute jobs will be updated when the user clicks the 'Refresh List' button located on that page.
- For multiple compute jobs: View the progress by clicking on the GUID, then checking the 'Compute Job Status' column.
The status of a compute job is defined by the following stages:
- Job Started: Indicates that the compute job has begun. This means that the compute job is preparing to run computation on the input dataset.
- Job Running: Indicates that a compute job is in progress and is running an algorithm to derive analytical insights from the input dataset. This could mean that the Compute Job is in the process of running the algorithm on the dataset, provisioning the data and/or algorithm, processing and filtering the results, and publishing the output file.
- Job Completed: Indicates a compute job is finished and the output file has been generated. The Data Consumer can proceed to download the output file.
- Job Failed: Indicates an unsuccessful compute job due to reasons such as failure in data provisioning, failure in algorithm provisioning, etc.
You can use these above-mentioned filters to refine your search of datasets based on the compute job status.
Viewing Log Files
To get a detailed record of time-stamped events and activity during the compute job process, you can view the Log Files generated after each compute job completion. You can download the log file from under the 'Compute Output' column by clicking on 'Get Result' and selecting the Log file to download.