How to investigate a job

Find the job

Open Jobs in the Console.

Use one or more filters:

• Result
• Runner set
• Repository
• Workflow
• Time range

The time range can be one hour, 24 hours, 7 days, or 30 days. Each page contains up to 50 jobs. Use Previous and Next to move through stored history.

If the table says No jobs recorded, confirm that a configured runner has accepted at least one job.

If the table says Job history unavailable, check the daemon log and confirm that the configured database path is writable.

Open job detail

Select a row to open the job detail panel.

Check:

1. Repository and workflow
2. Runner and backend
3. Queued time
4. Scale set assignment time
5. Runner assignment time
6. Start and completion time

These times show where a job spent time before its runner started work.

Read the runner log

Use Runner log to inspect runner process output.

For a running job, the Console requests new log chunks every two seconds. A completed job stops polling after all stored chunks are read.

If the panel says No log data is available, the backend did not collect a runner log for that job. Check the daemon log for backend startup or collection errors.

Inspect resource history

For a Docker job, the charts show container data:

• CPU
• Memory
• Disk read and write
• Network receive and send

For a Tart job without guest resource data, the panel shows Tart host estimate. Memory and disk values describe VM allocation. Host side values are estimates and do not describe exact guest use.

If a chart has fewer than two samples, the Console reports that no history is available yet.

See History and Storage Reference for data source, accuracy, and retention rules.

Open GitHub Actions

Select Open in GitHub Actions to open the matching workflow run and job.

Use the Console log to inspect runner startup and local resource behavior. Use GitHub Actions for step output, annotations, artifacts, and workflow controls.