Question 230
A database team repairs this problem: A sensor service needs high sustained write throughput and fast reads by device and time, without SQL joins. Which measured outcome validates the correction?
Explanation
Correct Answer
A realistic ingestion and range-scan test meets throughput and tail latency goals. The result directly checks whether the remedy worked: Evaluate Bigtable with row keys designed for device and time range reads. The cause was a relational query model was adopted even though access is dominated by key-ordered time-series retrieval.
Incorrect Answers
Even if this observation were recorded, it would support a case where a ticketing service running on Cloud SQL for PostgreSQL rejects new sessions during a marketing event while CPU remains moderate. It would not verify that a realistic ingestion and range-scan test meets throughput and tail latency goals.
Even if this observation were recorded, it would support a case where a high-throughput PostgreSQL-compatible analytics-heavy transactional application needs a managed service optimized for demanding reads. It would not verify that a realistic ingestion and range-scan test meets throughput and tail latency goals.
Even if this observation were recorded, it would support a case where a commerce application needs a relational database with foreign keys and joins, but architects propose Bigtable for every record. It would not verify that a realistic ingestion and range-scan test meets throughput and tail latency goals.