Practice Exam

Explanations for each answer:

• 3 is incorrect. While 3 matches the number of instances per App Service Plan, this doesn't align with the data structure or scaling needs. The data is organized by highway (6 highways), not by instance count. With only 3 partitions, you'd have suboptimal parallelism since multiple highways would share partitions, reducing throughput and increasing latency when processing time-series data for different highways.
• 4 is incorrect. While 4 matches the number of departments consuming the data, Event Hub partitions should be based on the logical partitioning of the input data, not the number of consumers. Multiple consumer groups can read from the same set of partitions. With 4 partitions for 6 highways, some highways would share partitions, reducing the parallelism available for data ingestion and processing.
• 6 is correct. Six partitions match the number of highways being monitored, providing the optimal data structure for maximum throughput and minimum latency. By using the highway identifier as the partition key, all traffic data for a given highway goes to a specific partition, ensuring ordered processing within each highway. This allows parallel processing of all six highways simultaneously and enables each WebJob to efficiently process data for specific highways without cross-partition queries.
• 12 is incorrect. While 12 partitions (4 departments × 3 instances) would work, it's excessive for this scenario and would reduce efficiency. Event Hub partition counts should match the logical data structure and scale needs. With only 6 highways generating data, having 12 partitions means half would be underutilized. Over-partitioning increases management overhead without providing throughput benefits when the partition key (highway) has only 6 unique values.