{# Total Spend #}

Total Spend

Total BigQuery execution cost from INFORMATION_SCHEMA.JOBS across the cached scan window.

across {{ "{:,}".format(totals.total_jobs) }} job{{ "s" if totals.total_jobs != 1 else "" }}

{# Build Spend — dbt-originated CTAS / MERGE / INSERT / UPDATE / DELETE #}

Build Spend

Spend on jobs classified as build — dbt-originated queries that materialise data (CTAS, MERGE, INSERT, UPDATE, DELETE). These are your warehouse build costs.

{% if totals.total_cost_usd and totals.build_spend %}

{{ "%.0f" | format(totals.build_spend / totals.total_cost_usd * 100) }}% of cached spend

{% else %}

dbt-originated writes

{% endif %}

{# Consumption Spend — non-dbt SELECT (BI tools, ad-hoc analysts) #}

Consumption Spend

Spend on jobs classified as consumption — non-dbt SELECT queries from BI tools, ad-hoc analysts, reverse-ETL exports, etc. These read your warehouse without writing to it.

{% if totals.total_cost_usd and totals.consumption_spend %}

{{ "%.0f" | format(totals.consumption_spend / totals.total_cost_usd * 100) }}% of cached spend

{% else %}

non-dbt readers

{% endif %}

{# Flagged Spend — measured cost of jobs touched by ≥1 detection rule. #}

Flagged Spend

Sum of current_cost across every active opportunity — i.e., money you actually spent on jobs that hit at least one detection rule. Savings projections aren't shown because governor-audit doesn't run shadow validation, so a heuristic percentage would be guess-work.

{% if totals.total_cost_usd and totals.flagged_spend %}

{{ "%.0f" | format(totals.flagged_spend / totals.total_cost_usd * 100) }}% of cached spend

{% else %}

spend on flagged jobs

{% endif %}