Master Pipeline - Payload Preparation: Executes Steps 2, and 3 sequentially with dependency checks between each Notebook activity.
Master Pipeline: Triggered manually by a user-initiated run or on a scheduled cadence. Rating Execution: Manages Step 4, 5 (Rating Engine calls) with parallel Spark job submission and concurrency throttling.
Result Processing: Executes Steps 6, 7, and 8 sequentially, with data quality gate checks between steps.
Serve & Refresh: Triggers Power BI dataset refresh (Step 9)
On-demand: Triggered by analyst via the What-If UI (Step 1) for immediate scenario execution.
Scheduled batch: Nightly execution of standing scenarios defined by the pricing team - typically run off-peak to avoid Rating Engine contention.(Step 2, 3)