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 Stream Analytics

  Why this is correct

  Azure Stream Analytics is a fully managed, serverless stream-processing engine that executes continuous, SQL-based queries over unbounded data. It provides built-in temporal windowing functions (tumbling, hopping, and sliding windows) that are ideal for analyzing clickstream events in near real time, and it can write results directly to Azure SQL Database or Power BI with minimal configuration. Unlike a batch tool, it maintains long-running, stateful queries with sub-minute latency, and its per-streaming-unit cost model makes it the simplest and most cost-effective choice for this retail clickstream scenario.

• ✗

  Azure Data Factory

  Why it's wrong here

  Azure Data Factory is designed for batch-oriented data integration and orchestration, using pipelines to copy, transform, and load data on scheduled triggers (for example, hourly or daily). Its activities execute finite workloads that read bounded datasets, so it does not offer a continuous query engine capable of processing an infinite event stream like live clickstream data. To handle streaming with ADF you would need to add a separate streaming service upstream, which makes it the wrong direct answer for real-time clickstream processing.

• ✗

  Azure Synapse Pipelines

  Why it's wrong here

  Azure Synapse Pipelines are built on the same engine as Azure Data Factory, so they inherit the same batch-centric design: they orchestrate flows that run on a fixed schedule or in response to discrete events, acting on bounded datasets rather than processing data continuously. Synapse Pipelines do not provide built-in streaming windows, event-time handling, or stateful aggregation that a live clickstream feed requires. Consequently, while Synapse can move and transform stored clickstream aggregates, it is not the appropriate service for the real-time streaming portion of this scenario.

• ✗

  Azure Databricks

  Why it's wrong here

  Azure Databricks does support real-time ingestion and Structured Streaming through Spark's micro-batch processing, but using it for this task requires provisioning and managing clusters, writing PySpark or Scala code, and explicitly implementing checkpointing and output sinks. This operational overhead, along with per-hour cluster costs, makes Databricks far more complex and expensive than a purpose-built, serverless SQL-based streaming service for the simple task of aggregating clickstream data. Therefore, while it is technically feasible, it is not the simplest or most cost-effective choice, which is why Stream Analytics is the correct answer.

Go deeper

Related to this question