Data warehouse FinOps connects platform usage to owners, budgets, and workload outcomes. The framework applies to compute, storage, data transfer, and managed services.
Collect daily usage, query or job activity, storage, transfer, and service-level charges. Keep a baseline long enough to include scheduled and peak workloads.
Assign usage to teams, projects, pipelines, and environments with tags, query metadata, or dedicated resources. Shared resources require an explicit allocation method.
Start with the largest measurable source of waste: idle compute, oversized resources, repeated scans, or unnecessary refreshes. Compare credits and runtime before and after each change.
Set owners, budgets, alerts, and review intervals. A monthly review should record the baseline, changes, savings, and performance effect. FinOps is an operating process, not a one-time tuning exercise.