One place for your AWS and GCP spend

For AWS, GetFinOps calls Cost Explorer in each account through the cross-account role you deploy, or reads the cost snapshot that the in-account collector writes to your bucket. A scan reads month-to-date cost, grouped by service and region for each day.

For GCP, it queries the Cloud Billing export in the BigQuery dataset you name at setup, with that project’s service account key. It checks whether you export standard or detailed usage cost and reads one table, never both, so no cost is counted twice.

The cost report shows the account selected in the header, or your primary account when none is selected. For GCP projects and AWS accounts on the cross-account role, it queries the cloud each time you open it. On those AWS accounts you get:

• Unblended or amortized cost, in daily or monthly points.
• A total, or a breakdown by service or by region, with a filter for the services you care about.
• Preset windows, month to date, or a custom range of up to 366 days.
• Tax, credit, refund and bundled discount rows excluded unless you tick “Incl. tax”.
• Cost Explorer’s forecast with an 80% prediction interval.

GCP accounts show daily cost by service from the billing export. The metric, granularity and tax toggles don’t apply to them, and a custom date range is read as a window of the same length ending today.

On daily views, GetFinOps works out a 7-day moving average, measures how far each day sits above it, and flags any day more than two standard deviations above the average. It needs at least seven days with an average before it flags anything.

Days below the average don’t widen the spread when enough days sit above it, so one large credit or refund can’t mask a real spike. Monthly points use a three-point average and are not flagged. The same check runs on AWS and GCP accounts.

Each scan stores cost for each connected account, plus a combined row for the whole workspace. The Cost by account card splits month-to-date spend across accounts. The Cost trend by account chart stacks each account’s daily spend and switches between dollars and share of the total.

Both charts appear once two or more accounts have cost data. The same per-account figures are available to AI assistants through the MCP server’s get_cost_by_account tool, and the cost report through get_cost_trend.

On AWS, deploy the cross-account role template or the in-account collector stack. The role’s read permissions include Cost Explorer cost and forecast access. Write access is a separate option that only remediation needs.

On GCP, turn on Cloud Billing export to BigQuery, create a service account with BigQuery Data Viewer on the billing export dataset and BigQuery Job User on the project, and enter the project ID, the billing export dataset ID and the JSON key when you connect.

• It doesn’t cover Azure or any cloud other than AWS and GCP.
• It doesn’t forecast GCP spend, or spend for AWS accounts that use the in-account collector.
• It doesn’t query AWS accounts on the in-account collector when you open the cost report. It shows the primary account’s latest month-to-date snapshot, where the metric, granularity, tax and date-range controls don’t apply, and nothing for other accounts on that collector.
• It doesn’t subtract GCP credits: GCP figures sum the billing export’s cost column.
• Spike flags mark days on the chart and count them. They don’t create findings or send alerts.
• The cost report shows one account at a time, not a sum of all accounts. Use the by-account charts for the whole workspace.
• The by-account charts update when a scan runs, not in real time.