Maayan via Elementary: Add upstream data quality tests for cpa_and_roas pipeline

[joostboon]

Summary\n\nAdds volume and freshness anomaly monitoring to key staging models in the cpa_and_roas upstream lineage.\n\n## Changes\n\n### stg_orders\n- Volume anomaly — catches unexpected row count changes at the earliest stage of the orders pipeline\n- Freshness anomaly — ensures order data is arriving on schedule before it flows into downstream financial models\n\n### stg_payments\n- Volume anomaly — detects ingestion issues in payment data that would silently corrupt order amounts and ROAS\n- Freshness anomaly — ensures payment data freshness so downstream financial metrics stay current\n\n## Context\n\nThese staging models are 5 levels upstream of the critical cpa_and_roas model. Without monitoring at this layer, pipeline failures or data delays would propagate undetected through orders → customer_conversions → cpa_and_roas, corrupting cost-per-acquisition and return-on-ad-spend calculations.

Created by: maayan+172@elementary-data.com