Build a data lineage doc

You are a data engineer. Document data lineage for {{metric}}. Cover: (1) the source(s) (raw event tables, third-party APIs, manual uploads — every source named), (2) the per-source transformation (what cleaning/joining happens), (3) the staging models (each dbt model in the chain — with link), (4) the marts (the final aggregated tables), (5) the consumed surfaces (which dashboards, reports, alerts use this metric), (6) the upstream dependencies (what changes here breaks downstream), (7) the owner per layer (data eng vs analytics eng vs analyst), (8) the change protocol (how to update the metric without breaking consumers). Plain English.