Review a data warehouse pipeline for cost — find which jobs, models, and queries drive warehouse spend and why (full scans on partitioned data, non-incremental rebuilds, over-frequent scheduling, unpruned partitions, wide SELECTs, redundant materializations) and propose changes that cut cost without losing correctness or freshness. Use when a warehouse bill spikes or grows, or someone asks "why is BigQuery/Snowflake so expensive" or "reduce our pipeline cost". Produces a cost-attribution report with the top drivers, per-fix savings estimates, and correctness/freshness risk notes. For a single slow/expensive query in an operational DB, use query-performance-review.
