Connect data sources
Read billing from AWS, Azure and Google Cloud, add Kubernetes allocation from OpenCost, and upload or push anything else as FOCUS.
On this page
Sources
| Source | What you connect | The checklist confirms |
|---|---|---|
| AWS | A Data Export in S3, read through a role with an external ID. FOCUS, or CUR 2.0 which is converted | It can list the export and read a file |
| Azure | A Cost Management FOCUS export in Blob Storage | The container is set and the export can be listed |
| Google Cloud | A FOCUS view over the billing export in BigQuery | Project and dataset are set and the view can be queried |
| OpenCost | The OpenCost URL and cluster name | It can read yesterday's allocation |
Upload or push
- Upload a FOCUS file as CSV, gzipped CSV or Parquet. Oracle Cloud's FOCUS report and SaaS invoices arrive this way.
- Push rows through the API with a key from Settings. UnitVane AI token spend arrives this way.
How loads work
- FOCUS 1.0 to 1.4 is accepted.
- Each load replaces the billing months it contains, so loading a month twice still counts it once.
- A provider restatement replaces the month whole.
- Every run is listed with its rows, billing periods and FOCUS version.
Kubernetes
OpenCost weights split each cluster's billed cost across namespaces, so the split always adds up to what the cloud charged. For pod-level allocation and change proposals, see UnitVane.
Need help with this guide?Contact support