DP-900 Describe an analytics workload on Azure Practice Question

Answer analysis

Option-by-option breakdown

For each option: why learners choose it and why it is or isn't the right answer here.

• ✗

  Azure Functions

  Why it's wrong here

  Azure Functions is a serverless compute service designed to execute short-lived, event-driven code in response to a trigger, such as a single Event Hubs message or batch. It has no built-in temporal windowing, stateful stream aggregation, or SQL-like query engine, so continuously aggregating thousands of IoT device streams would require manually implementing window state and exactly-once semantics in code. That makes it appropriate for per-event reactions like alerting or enrichment, not sustained low-latency analytical aggregates.

• ✗

  Azure Databricks

  Why it's wrong here

  Azure Databricks is a powerful analytics platform for complex data engineering, machine learning, and data science, including processing streaming data with Spark. However, for straightforward real-time transformations and aggregations, it introduces the overhead of managing Spark clusters, which is not ideal for a simple, continuous, low-latency pipeline. It would be the correct choice if the scenario required advanced analytics, machine learning inference, or highly complex, stateful stream processing beyond basic aggregations.

• ✓

  Azure Stream Analytics

  Why this is correct

  Azure Stream Analytics is a fully managed stream-processing engine built precisely for real-time, high-throughput IoT telemetry. It reads directly from Event Hubs or IoT Hub and runs declarative SQL-like queries that support tumbling, hopping, sliding, and session windows for continuous in-memory aggregation. Because the service handles checkpointing, scaling, and fault tolerance automatically, it delivers consistent sub-second results without the operational burden of managing clusters or writing custom stateful code.

• ✗

  Azure Data Factory

  Why it's wrong here

  Azure Data Factory is a cloud data integration and orchestration service centered on scheduled pipelines, batch ETL, and copy activities across stores. It lacks an engine for unbounded streams and time-windowed queries, so feeding it thousands of live device events would mean either creating stop-gap micro-batches with noticeable latency or wiring it to another streaming service first. Its natural role is periodic batch movement, not always-on real-time aggregation.

Go deeper

Related to this question