> For the complete documentation index, see [llms.txt](https://docs.onum.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.onum.com/data-sinks-changelog/amazon-s3-data-sink.md).

# Amazon S3 Data Sink

Learn more about this Data Sink type in [this article](/data-sinks/data-sink-integrations/send-data-to-amazon-s3.md).

<details>

<summary>v2.1.0</summary>

<mark style="background-color:purple;">**Released on**</mark> <mark style="background-color:purple;"></mark><mark style="background-color:purple;">05/13/2026</mark>

* New **Assume Role** support via STS. This allows users to assume an IAM role at runtime using a Role ARN, External ID, role session name, STS region, and session duration. Useful for cross-account access or restricted permission boundaries.
* **Auto authentication mode**: Automatically uses the default AWS credentials available in the environment (instance profile, ECS task role, environment variables, etc.) without requiring explicit access keys.

</details>

<details>

<summary>v2.0.0</summary>

<mark style="background-color:purple;">**Released on**</mark> <mark style="background-color:purple;"></mark><mark style="background-color:purple;">02/18/2026</mark>

* Added proxy configuration support for HTTP and HTTPs proxies.
* Added proxy authentication with username and password.
* Added exception list to bypass proxy for specific hosts.

</details>

<details>

<summary>v1.0.1</summary>

<mark style="background-color:purple;">**Released on**</mark> <mark style="background-color:purple;"></mark><mark style="background-color:purple;">08/27/2025</mark>

* Fixed race condition when uploading Parquet files.

</details>

<details>

<summary>v1.0.0</summary>

<mark style="background-color:purple;">**Released on**</mark> <mark style="background-color:purple;"></mark><mark style="background-color:purple;">06/02/2025</mark>

* The raw/parquet configuration can now be changed per Pipeline.
* Prefixes are now configurable per Pipeline.
* A global prefix is now added so all paths for a Data sink have this common prefix.

</details>

<details>

<summary>v0.2.0</summary>

<mark style="background-color:purple;">**Released on**</mark> <mark style="background-color:purple;"></mark><mark style="background-color:purple;">05/14/2025</mark>

* Real time is used to calculate object prefixes and deadlines now, instead of event timestamp or Pipeline start time.
* Now, the Data Sink uploads pending objects when their deadline is reached, even without traffic.

</details>

<details>

<summary>v0.1.0</summary>

<mark style="background-color:purple;">**Released on**</mark> <mark style="background-color:purple;"></mark><mark style="background-color:purple;">02/18/2025</mark>

* Added parquet format option for S3 objects.

</details>

<details>

<summary>v0.0.3</summary>

<mark style="background-color:purple;">**Released on**</mark> <mark style="background-color:purple;"></mark><mark style="background-color:purple;">02/10/2025</mark>

* Fixed the `EntityTooSmall` error while uploading to S3.
* Fixed sending an empty part list to `CompleteMultipartUpload`.

</details>

<details>

<summary>v0.0.2</summary>

<mark style="background-color:purple;">**Released on**</mark> <mark style="background-color:purple;"></mark><mark style="background-color:purple;">02/28/2024</mark>

General fixes

</details>

<details>

<summary>v0.0.1</summary>

<mark style="background-color:purple;">**Released on**</mark> <mark style="background-color:purple;"></mark><mark style="background-color:purple;">01/31/2024</mark>

Initial version

</details>