AWS Certified Developer Associate DVA-C02 Practice Question

When a database such as Amazon DynamoDB returns a ProvisionedThroughputExceededException, the request rate has temporarily exceeded the allocated write capacity. The recommended best practice is to retry the operation using exponential backoff and jitter-each retry waits progressively longer, with a small random variation, so that concurrent clients are less likely to retry in unison. This gives the table time to recover and spreads load more evenly.

Retrying at a fixed interval keeps applying pressure at the same frequency and is more likely to continue failing. Dynamically changing provisioned capacity from inside the application adds latency, may not take effect quickly enough, and raises cost. Switching the table to on-demand capacity requires an UpdateTable call that cannot be toggled frequently and can only be performed once every 24 hours, so it is unsuitable for per-request error handling.