Provisioned throughput is the maximum read and write capacity that an application can use within a table or index. 

If we use more than the specified RCU or WCU, DynamoDB can throttle the requests, the requests will fail with a 400 (bad request) error, and DynamoDB throws ProvisionedThroughputExceededException. AWS SDKs can do automatic retries in the case of a throughput exception. 

Apart from manual throughput provisioning, DynamoDB also supports features such as auto-scaling and reserved capacity. Having a decent understanding of throughput provisioning can help you configure these alternate options efficiently. We will discuss different throughput provisioning scenarios in this recipe.