Which service should you use to process a stream of sensor data and aggregate values over one-minute windows?

Using Azure Stream Analytics is ideal for processing a stream of sensor data and aggregating values over specified time windows, such as one-minute intervals. This real-time analytics service is designed specifically for working with streaming data, enabling near real-time insights from large data flows.

Azure Stream Analytics offers features such as time windowing, where you can define how data is grouped and aggregated over specified periods. Its SQL-like query language allows for straightforward expressions to perform aggregations, such as sum, average, or count, over specified windows of time. This makes it particularly suitable for scenarios involving continuous data streams, such as sensor data.

While Azure SQL Database provides robust relational database capabilities, it is more suited for structured data and does not inherently support real-time stream processing. Azure Cosmos DB excels in handling large amounts of distributed data and offers low-latency access but is not primarily designed for streaming analytics. Azure Data Factory is primarily focused on orchestration and data integration, allowing for the creation and management of data workflows but does not directly handle real-time stream analytics like Azure Stream Analytics does.

Thus, Azure Stream Analytics is the appropriate choice for aggregating sensor data over defined time intervals, making it the best fit for this requirement.