Question

A developer is designing a DynamoDB table that ingests high-volume IoT sensor readings. All devices report using the same 'metric-type' value (temperature), which the team initially chose as the partition key. During load tests, they observe severe write throttling on a single partition even though total provisioned write capacity is far from being fully consumed. Which change to the table design will best distribute the write load?

Accepted Answer

Add a random suffix (e.g., a value from 1–10) to the partition key value to create a write-sharding scheme, and store the original metric-type as a separate attribute.

Answer

Switch the table to on-demand capacity mode so DynamoDB automatically avoids all throttling regardless of key distribution.

Answer

Create a global secondary index on the timestamp attribute so writes are routed to the index instead of the base table.

Answer

Increase the provisioned write capacity units on the table until the throttling stops during the load test.