How should this exception be handled?
Create a new global secondary index for the table to help with the additional requests.
Retry the failed read requests with exponential backoff.
Immediately retry the failed read requests.
Use the DynamoDB “UpdateItem” API to increase the provisioned throughput capacity of the table.
Explanations:
Adding a new global secondary index (GSI) does not directly address the issue of ProvisionedThroughputExceeded errors caused by request bursts. A GSI can help with read capacity if the query patterns change, but it does not solve the issue of throughput exceeded due to traffic spikes.
Exponential backoff is the recommended way to handle throttling errors like ProvisionedThroughputExceeded. It helps to avoid overloading the system by increasing the time between retries, giving DynamoDB time to recover from the burst in traffic.
Immediately retrying without backoff can worsen the problem, as it may lead to repeated throughput exhaustion. This approach can result in further throttling and delays.
Increasing the provisioned throughput capacity may help in the long run, but it is not an immediate solution to the intermittent bursts of traffic. Also, this option is more about scaling the table’s capacity rather than dealing with a specific error scenario.
I deduce that the answer is:
Retry the failed read requests with exponential backoff.