Question

A developer builds a Python application using the AWS SDK (boto3) to write items to a DynamoDB table. During traffic spikes, the application occasionally receives ProvisionedThroughputExceededException and other transient 5xx errors, causing some writes to fail permanently. The developer wants the SDK to automatically handle these transient failures with the least amount of custom code. What is the recommended approach?

Accepted Answer

Configure the SDK client's retry settings (e.g., adaptive or standard retry mode with a suitable max attempts value), which applies exponential backoff with jitter to retryable errors automatically.

Answer

Wrap every DynamoDB call in a try/except loop that immediately retries the request a fixed number of times with no delay between attempts.

Answer

Switch all writes to a synchronous invocation model and increase the client socket timeout so requests wait longer before failing.

Answer

Catch the exceptions and route the failed items to an SQS dead-letter queue, then manually reprocess them from a separate script.