Monitor Data Storage and Data Processing DP-203 Practice Questions (Page 7)

Questions 31–35

1. 31application · medium

   A company runs a real-time stream processing job on Azure Stream Analytics that ingests IoT telemetry. The job's 'Watermark delay' metric has been increasing over the past hour, and the 'Input events' metric is flat. The team suspects the job is falling behind. What should they do first to diagnose the issue?

   • ACheck the job's 'Runtime errors' metric and then modify the query to use a larger window size.
   • BIncrease the number of streaming units without reviewing any metrics.
   • CReview the job's 'Watermark delay' and 'Input events' metrics in Azure Monitor to confirm the backlog, then check the 'CPU % utilization' metric for the streaming nodes.
   • DRestart the Stream Analytics job to reset the watermark delay and clear the backlog.
   Reveal Solution Discussion Select an answer first
2. 32expert · hard

   A data engineer is troubleshooting a slow query in Azure Synapse Analytics. The query joins a large fact table with a small dimension table. The execution plan shows a broadcast join is being used, but the fact table is large. The engineer suspects the statistics on the dimension table are stale, causing the optimizer to choose a broadcast join. What should they do to confirm and fix the issue?

   • AIncrease the memory for the query to make the broadcast join more efficient.
   • BRebuild the dimension table to improve its statistics.
   • CUpdate statistics on the dimension table and re-run the query to see if the execution plan changes to a shuffle join.
   • DForce a shuffle join by using a query hint to avoid the broadcast join.
   Reveal Solution Discussion Select an answer first
4. 33expert · hard

   A data team uses Azure Data Factory to orchestrate a complex pipeline that loads multiple tables. They want to run a data validation test after each pipeline run and automatically rerun the pipeline if the test fails due to transient issues, but not if the test fails due to data quality issues. What should they implement?

   • AAdd a validation activity at the end of the pipeline that checks row counts, and use a retry policy on the validation activity.
   • BUse an Azure Function activity to run the validation, and if it fails, check the error message to decide whether to trigger a rerun of the pipeline.
   • CCreate a separate validation pipeline and run it after the main pipeline using a trigger with a dependency.
   • DUse a 'Until' activity that loops the entire pipeline until the validation passes.
   Reveal Solution Discussion Select an answer first
5. 34application · medium

   A company needs to track all data processing activities in Azure Databricks for audit purposes. They want to capture who ran what notebook, when, and the outcome. What should they configure?

   • AEnable diagnostic settings on the Azure Databricks workspace to send logs to a Log Analytics workspace.
   • BTurn on Azure Monitor for VMs on the Databricks cluster nodes.
   • CConfigure a Log Analytics agent on each cluster node to collect custom logs.
   • DUse Azure Activity Log to capture all Databricks workspace operations.
   Reveal Solution Discussion Select an answer first
6. 35application · medium

   A data engineering team has a critical pipeline in Azure Data Factory that loads a sales data mart. They want to validate the pipeline's functionality by running a test after every deployment and monitoring the test results. What should they implement?

   • ACreate an Azure DevOps pipeline that runs a data validation script after the Data Factory deployment and publishes the results to a Log Analytics workspace.
   • BUse Azure Monitor alerts on the pipeline's 'Failed Runs' metric and assume the pipeline is working if no alerts fire.
   • CUse Azure Data Factory's 'Trigger' with a scheduled time to run the pipeline and then check the pipeline run status in the monitoring view.
   • DManually run the pipeline after each deployment and record the results in a shared spreadsheet.
   Reveal Solution Discussion Select an answer first