Results & reporting
Monitor an evaluation
Track evaluation progress and confirm when the result is ready to review.
Use the evaluation record to follow progress, identify execution problems, and confirm that scoring has finished. Completion means the workflow ended; it does not mean the target passed.
Read the run state
| State | Meaning | What to check |
|---|---|---|
| Queued | The evaluation is waiting to start. | Confirm the target and configuration are still valid. |
| Running | Test cases are being sent and responses are being collected. | Watch completed cases, errors, retries, and stalled progress. |
| Scoring | Responses are being evaluated and aggregated. | Wait until scoring and finalization finish. |
| Completed | Execution and scoring finished. | Confirm the final result and its evidence coverage. |
| Failed | The workflow stopped before valid completion. | Review the cause; do not treat it as a completed result. |
Completion checklist
- The evaluation state is Completed, not Failed.
- The target and evaluation versions match the intended scope.
- Scoring is complete and the displayed result is final.
- Failed, excluded, or unresolved cases are accounted for.