A ticketing system stores all open-ticket records in one DynamoDB table whose partition key is the status field, which can only be open, pending, or closed. CloudWatch shows throttling on the table during peak support hours. Which fix addresses the root cause?
Show answer and explanation
Answer: B
With only three status values, nearly all writes land on three logical keys, and a single partition key value is limited to roughly 1,000 write units per second regardless of how much table capacity is provisioned. Choosing a high-cardinality key such as the ticket identifier spreads writes across partitions, and write sharding with a suffix helps when a status-based access pattern must remain. Adaptive capacity helps only partly.
Why the other options are wrong
- A. Adding table capacity cannot push past the per-partition throughput limit for one key value, so the throttling continues even with a larger bill.
- C. Disabling streams only removes downstream change capture and leaves the write hotspot exactly as it was, so throttling on the table persists.
- D. Dropping writes silently loses ticket data, which is worse than throttling, and it does nothing to improve how the key distributes load.