Cost insights setup
StaleSweep can read your AWS Cost & Usage Report (CUR) to show real spend — a bill breakdown by service, region, and account, cost-anomaly alerts, and the actual cost of each unused resource it finds.
Unlike tools that poll the Cost Explorer API (which AWS bills at $0.01 per request), StaleSweep reads the CUR delivered to S3 — so the ongoing AWS cost to you is effectively zero (a few cents of S3 storage). Cost data stays in your account; StaleSweep reads it with the same read-only scan role you already connected, and stores only daily aggregates.
EnableCostReports=true), deployed in us-east-1 — the only region AWS allows the CUR export. The scan role is global IAM, so a us-east-1 stack still scans every region; one stack creates the role and the CUR export together.Deploy the stack in us-east-1 (recommended: auto bucket)
The Deploy cost insights button on the Spend page opens CloudFormation pinned to us-east-1 with cost reports pre-enabled. It provisions an S3 bucket and a CUR 2.0 export, and grants the scan role read access — one deploy.
- On the Spend page, click Deploy cost insights (us-east-1). It opens the connection stack with
EnableCostReports=trueand the region prefilled. - Leave
CostReportsBucketas prefilled (a uniquestalesweep-<random>name we propose for you) so the stack creates a dedicated bucket and delivers the CUR there. - Create the stack. That's it — role, bucket, and export in one go.
Bring your own bucket (optional)
Set CostReportsBucket to a bucket you own and the stack delivers the CUR there instead — it still creates the export and attaches a delivery policy to that bucket. The bucket must already exist, and its policy is replaced, so use one dedicated to this CUR.
What to expect after enabling
- AWS can take up to 24 hours to deliver the first CUR. The Spendpage shows a “waiting for your first report” state until it lands, then fills in automatically.
- Once data arrives, Spend shows your month-to-date total — compared against the same period last month, not a rolling 30-day window — plus a breakdown you can group by service / region / account / usage type, and a banner for any detected cost spikes.
- Each unused-resource finding then shows what that specific resource actually cost — so you can prioritize cleanup by real dollars, not estimates.
- The Dashboard's savings card also shows your estimated waste as a share of that month's bill, projected to a full-month estimate from your pace so far — a sense of scale for how much of what you're paying AWS is avoidable.
