Job Runs

Each invocation of a Job template produces a Job Run with introspection capabilities

Triggering Job Runs

Jobs Runs can be triggered manually through the Overview > Jobs page and pressing Trigger for a specific Job.

Inspecting Job Runs

Details about Job Runs can be visualized either through Overview > Jobs > Inspector or by visiting Inspector on the left-hand sidebar and clicking on the Job entity. Runs are presented in reverse chronological order. This view exposes a number of filtering options.

  • Name: This dropdown allows for filtering the displayed Job Runs by their user-friendly names.
  • Job Run ID: This filter allows for finding the results for a specific Job Run ID. This will be of the UUIDv4 format.
  • Status: This filter allows for displaying Job Runs based on their current status.

For a give Job Run, the logs, resource utilization metrics, and cost dashboards are provided in this view. These are presented using the Kaspian-managed Grafana dashboard hosted within the user’s compute cluster.

Logs

The following is an example of how the Grafana logs dashboard looks for Job Runs. It is primarily an chronological log stream. Due to limitations of Grafana, by default a given time slice is limited to 5000 lines of logs.

Metrics

The following is an example of how the Grafana metrics dashboard looks for Job Runs. This view presents various statistics, including CPU utilization, RAM utilization, filesystem I/O, networking I/O, etc.

Cost

The following is an example of how the Grafana cost dashboard looks for Job Runs.