Question about retrying batch writes in DynamoDB using C#

[bykka]

Hi,

I have a question regarding the behavior of the DynamoDB client for .NET, specifically its handling of retries and exceptions during batch write operations.

According to the documentation, the DynamoDB client for .NET performs up to 10 retries by default for requests that fail due to server-side throttling. However, the batch write API documentation does not explicitly describe the potential errors or exceptions that could be thrown during its operation.

If I have a table with low provisioned capacity and I perform a massive update operation using the batch write API, is it possible for some writes to fail silently (i.e., not get saved) without the client throwing an exception or providing a clear indication of the failure?

If so, how can I reliably detect and handle such cases to ensure data consistency?