An application generates log events at 10 Hz, each event containing five different metrics, producing high data volume. The application will write logs to Amazon Timestream and run a daily query against the table. Which combination of steps gives the fastest query performance? (Choose three.)

Batch writes improve performance by reducing the number of API calls and network overhead, which is efficient for high-volume data like log events. Writing each log event individually would be less efficient due to increased overhead. Storing each log as a multi-measure record is optimal because Timestream is designed for time-series data with multiple metrics per timestamp. This allows for efficient storage and querying of related metrics together. Storing each log as a single-measure record would be less efficient as it would require more rows for the same data, increasing query complexity. Setting the memory-store retention period shorter than the magnetic-store retention period ensures that frequently accessed recent data is in the faster memory store, while older, less frequently accessed data is moved to the more cost-effective magnetic store. This optimizes both performance and cost. Setting it longer would be inefficient as it would keep less frequently accessed data in the more expensive memory store.