How should the database administrator remediate this issue at the lowest cost?
Enable auto scaling and set the target usage rate to 90%.
Switch the table to provisioned mode and enable auto scaling.
Switch the table to provisioned mode and set the throughput to the peak value.
Create a DynamoDB Accelerator cluster and use it to access the DynamoDB table.
Explanations:
Enabling auto scaling with a target usage rate of 90% may not be sufficient to handle the sudden and large spike in traffic. The high concurrency around 5 PM can lead to throttling if DynamoDB doesn’t scale fast enough.
Switching to provisioned mode and enabling auto scaling allows the system to handle predictable spikes in traffic, as auto scaling can adjust read and write capacity in response to demand.
Switching to provisioned mode and setting the throughput to the peak value could be very costly, as the system would need to reserve enough capacity for the peak load every day. This is inefficient and expensive, especially with fluctuating demand.
DynamoDB Accelerator (DAX) provides caching to improve read performance but does not directly address throttling issues related to write operations, which is the primary problem in this case.
It seems to me that the answer is:
Switch the table to provisioned mode and enable auto scaling.