> ## Documentation Index
> Fetch the complete documentation index at: https://docs.costgraph.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Object Store

> Read a bill your provider already publishes to storage - S3, Azure Blob, or anything S3-compatible - and push it to CostGraph

Some providers do not answer a billing API; they write the bill to storage on a
schedule. AWS Data Exports writes Parquet to S3, Azure Cost Management writes to
Azure Blob Storage, and STACKIT publishes `part_000X.csv.gz` to its own Object
Storage. The Object Store integration reads any of them: the rows are already
FOCUS when they arrive, so nothing is normalized - the exporter reads, repairs
the deviations it knows about, and pushes.

This integration runs in **your** environment via the
[exporter](/costgraph/integrations/exporter) - your storage credentials never
reach CostGraph, and only the finished FOCUS records do.

## Where the bill can live

| Where the bill is                        | Set                                                                                                      |
| ---------------------------------------- | -------------------------------------------------------------------------------------------------------- |
| AWS S3                                   | `OBJECTSTORE_BUCKET`, and nothing else if the exporter's role can read it                                |
| Azure Blob Storage                       | `OBJECTSTORE_ENDPOINT=https://<account>.blob.core.windows.net` and the container in `OBJECTSTORE_BUCKET` |
| S3-compatible (STACKIT, R2, MinIO, Ceph) | `OBJECTSTORE_ENDPOINT` and a read-only key pair                                                          |
| Files you already have                   | `OBJECTSTORE_ENDPOINT=file:///bills`                                                                     |

`.csv`, `.csv.gz` and `.parquet` are all read. Credentials are read-only by
intent: the exporter never writes to your bucket, so give it a key that cannot
either.

## Which files make up a period

`OBJECTSTORE_LAYOUT` answers one question: given all the files in the bucket,
which ones belong to a billing period?

| Layout           | Reads                                        | Use it for                                                           |
| ---------------- | -------------------------------------------- | -------------------------------------------------------------------- |
| `flat` (default) | every file under the prefix                  | a bucket that holds one copy of each file                            |
| `latest-run`     | only the newest export of each period        | a provider that rewrites the whole period on every run, like STACKIT |
| `manifest`       | exactly the files a published manifest lists | AWS Data Exports and Azure, which write one naming their data files  |

Prefer `manifest` where the export publishes one: nothing is guessed from file
names, and a run that lands while the exporter is reading cannot be read
half-written.

## Connect

1. Open **Settings -> Integrations** and choose **Object Store**. Pick the
   tenant the charges belong to and copy the **connection id** (`fpc_...`).
2. Create an API key with the **`focus:write`** scope - and
   **`deployment:register`** if you will run the exporter's container.
3. Run the exporter where the bill is reachable:

```bash theme={null}
OBJECTSTORE_BUCKET=my-focus-export \
OBJECTSTORE_LAYOUT=manifest \
COSTGRAPH_API_KEY=... COSTGRAPH_CONNECTION_ID=fpc_... \
focus-exporter --provider objectstore --sink costgraph
```

[Run the exporter](/costgraph/integrations/exporter) covers the container, the
Helm chart, and the binary; the
[provider reference](/costgraph/integrations/exporter-reference) lists every
variable. Without `--month` the current month is read, which is what a schedule
wants; a month can be pushed again at any time - a push replaces the window it
covers rather than doubling it.

## What CostGraph does with it

Records land as raw billing rows, are normalised into line items, and roll up
into daily cost. One connection carries one provider's rows - the provider is
stamped on first push - so exporting bills from two providers means two
connections.