A FinOps analyst needs the top spenders by compute usage over the last week, broken down by user and by job, using BigQuery metadata. Which source provides per-job compute totals?
Show answer and explanation
Answer: C
The INFORMATION_SCHEMA JOBS_BY_PROJECT view lists jobs run in a project, including the user, creation time, and total_slot_ms, the total slot time each job consumed. Filtering to the last seven days and grouping by user or job gives the per-job compute picture the analyst needs. Invoices summarize spend without job detail, and project creation logs and dataset listings do not record the compute consumed by queries.
Why the other options are wrong
- A. Invoices show spend totals per project, not the individual jobs behind them. An analyst cannot identify which user or query consumed the compute from a monthly PDF.
- B. Audit logs for project creation record administrative events, not query compute. They would not show how many slot milliseconds any job consumed.
- D. Dataset listings describe datasets and their modification times. They contain no record of jobs, the users running them, or the compute consumed by queries.