Monitoring DATA-ENGINEER-PROFESSIONAL Practice Questions (Page 4)

Questions 16–20

1. 16foundation · easy

   Which component of the Query Profiler UI provides operator-level metrics such as the number of output rows and duration for each operation in a query plan?

   • AQuery text editor
   • BQuery plan visualization
   • CQuery history list
   • DCluster metrics panel
   Reveal Solution Discussion Select an answer first
2. 17foundation · easy

   When reviewing `system.access.audit` data, you notice a series of `login` events from a single user at unusual hours. What is the most appropriate next step?

   • ADisable multi-factor authentication for the user to simplify access.
   • BIgnore the events because they are just logins.
   • CImmediately revoke the user's access.
   • DInvestigate the events to determine if they are legitimate or a potential security concern.
   Reveal Solution Discussion Select an answer first
4. 18application · medium

   A data analyst reports that a frequently used SQL query in Databricks SQL is running slower than usual. The query joins several large tables and has been working fine until recently. You need to identify the specific operator that is causing the bottleneck. Which tool should you use?

   • AQuery the system.query.history table to see the query's total execution time and compare it with previous runs.
   • BUse the Spark UI to view the underlying Spark jobs and look at the stage-level metrics for the query.
   • CRun the Databricks CLI command 'databricks sql query-history list' to get a list of recent queries and their durations.
   • DOpen the Query Profiler UI for the query and examine the operator-level metrics, such as the number of output rows and duration for each operator.
   Reveal Solution Discussion Select an answer first
5. 19application · medium

   A data engineer wants to identify which clusters have been idle for more than 2 hours to potentially terminate them and save costs. They need to query system tables to find clusters with no active jobs or queries. Which query approach should they use?

   • AQuery system.jobs.runs to find clusters that have not run any jobs in the past 2 hours.
   • BUse the Spark UI to check the cluster's executor metrics and see if there is any activity.
   • CQuery system.compute.cluster_details to get cluster state and last activity time, then filter for clusters with state 'RUNNING' and last_activity older than 2 hours.
   • DQuery system.billing.usage to find clusters with zero usage in the past 2 hours.
   Reveal Solution Discussion Select an answer first
6. 20application · medium

   A data engineer is investigating a Spark job that is running much slower than expected. They suspect that the cluster is under-provisioned, but they want to confirm by looking at executor metrics such as memory usage and task concurrency. Which tool should they use?

   • ARun the Databricks CLI command 'databricks clusters get' to retrieve the cluster's current configuration and see the instance type.
   • BUse the Query Profiler UI to view the SQL query's operator-level metrics and see if any operator is spilling to disk.
   • CQuery the system.compute.cluster_details table to see the cluster's autoscaling metrics and determine if it scaled up during the job.
   • DOpen the Spark UI and navigate to the Executors tab to view metrics like memory used, task time, and number of tasks per executor.
   Reveal Solution Discussion Select an answer first