dc_001 · Deployment: daily score refresh (happy path)
Status: ✓ Passing · Last run: 2026-06-09T06:20:16 · Pipeline: deployment
· History: 1/2 runs passed (50%)
What was tested
A model is already deployed and has been scoring rows for weeks. Today, 100 new rows arrived. The scheduled fire should score exactly those 100, write them to the output table, and leave the 1000-row prior_predictions table untouched. The simplest deployment contract: same-schema new data → predictions appended, history not corrupted.
What we planted in the data
- 1,000 entities
- Signal strength: 0.50 (sigmoid slope multiplier)
n_new_rows=100n_prior=1000
What the system did
- Training rows built: 1,000
- Features used in model: 3
- Model selected:
DeploymentScoringRunner
What we checked — all assertions passed
| Status | Assertion | Detail |
|---|---|---|
| ✓ | deployment_completed_cleanly |
clean run |
| ✓ | scored_row_count_eq |
scored 100 rows, expected 100 |
| ✓ | prediction_distribution_in_band |
predictions in [0.001, 1.000] (mean 0.469) |
| ✓ | prior_predictions_unchanged |
prior predictions row count unchanged (1000) |
How this could have gone wrong
(no assertion descriptions provided)
Why this case matters
This is the most common production deployment scenario — happy-path daily refresh. If THIS regresses, no deployment is safe.
Reproducing
# from auto_insight_api/
python -m validation.v2 run dc_001 --pipeline deployment -v
- Case config:
validation/v2/cases/dc_001_score_refresh.yaml - Data shape:
deployment_score_refresh - Analytics type:
deployment