Skip to content

Data Analyst Associate · Domain 6: Working with Dashboards and Visualizations in Databricks · 16% of the exam

Task 6.3: Schedule an automatic dashboard refresh; Configure an alert with a desired threshold and destination

Dashboard schedules and subscriptions, and Databricks SQL alerts with their query, condition, threshold, schedule and notification destinations.

Study it

  • Dashboard schedules and subscriptions

    Lesson coming

  • Databricks SQL alerts: thresholds and destinations

    Partly covered by: CloudWatch, Alarms & X-Ray

Sample questions

Try each one before opening the answer. Every option is explained, with the Databricks documentation page that proves it.

Question 1 · choose 1

An analyst creates a dashboard schedule that runs every 5 hours, starting at 06:00. On the following day, at which times before noon does the schedule run?

  1. A02:00 and 07:00
  2. B06:00 and 11:00 only
  3. C02:00, 06:00 and 11:00
  4. D01:00, 06:00 and 11:00
Show the answer and why
  • A02:00 and 07:00

    Incorrect

    The 5-hour steps do not continue past the start time; the cadence resets at 06:00 each day.

  • B06:00 and 11:00 only

    Incorrect

    The cadence continues until 05:59 the next day, so a run at 02:00 also takes place.

  • C02:00, 06:00 and 11:00

    Correct

    Runs at 06:00, 11:00, 16:00, 21:00 and 02:00 continue until 05:59; the cadence then resets and starts again at 06:00, followed by 11:00.

  • D01:00, 06:00 and 11:00

    Incorrect

    The overnight run follows 21:00 by five hours, which is 02:00.

For schedules with a defined start time, the refresh cadence begins at that time each day and repeats until just before the start time the next day, when it resets. Every 5 hours from 06:00 gives 06:00, 11:00, 16:00, 21:00 and 02:00, then 06:00 again.

Question 2 · choose 1

Finance wants an email to the finance-ops group when two conditions hold together: more than 10,000 draft invoices and more than 5,000 archived invoices. One query counts both. How should the analyst set up the Databricks SQL alert?

  1. ACreate two alerts, one per count, each emailing finance-ops when its threshold is crossed
  2. BHave the query return CASE WHEN drafts_count > 10000 AND archived_count > 5000 THEN 1 ELSE 0 END, and alert when that value equals 1
  3. CKeep the two count columns and type both conditions joined with AND into the alert's threshold field
  4. DSet a notification frequency so that the alert repeats until both counts fall back below their limits
Show the answer and why
  • ACreate two alerts, one per count, each emailing finance-ops when its threshold is crossed

    Incorrect

    Each alert evaluates its own condition, so finance-ops would be emailed when either count is high, not only when both are.

  • BHave the query return CASE WHEN drafts_count > 10000 AND archived_count > 5000 THEN 1 ELSE 0 END, and alert when that value equals 1

    Correct

    Alert query patterns combine several conditions in a CASE expression so that the alert checks one value against its threshold.

  • CKeep the two count columns and type both conditions joined with AND into the alert's threshold field

    Incorrect

    The alert condition compares a value with a threshold; combining conditions is done in the query, for example with CASE.

  • DSet a notification frequency so that the alert repeats until both counts fall back below their limits

    Incorrect

    The notification frequency repeats notifications until the alert returns to OK; it does not combine conditions.

A Databricks SQL alert runs its query on a SQL warehouse on a schedule, checks a value against a threshold, and notifies users or notification destinations when the condition is met. To require several conditions at once, the query returns a single flag computed with a CASE expression.

Question 3 · choose 1

During a two-week data migration, an analyst wants a dashboard's schedule to stop running and stop sending emails, but keep its settings and subscribers for later. What should she do?

  1. ADelete the schedule and recreate it after the migration
  2. BRemove all subscribers from the schedule
  3. CUnpublish the dashboard
  4. DPause the schedule
Show the answer and why
  • ADelete the schedule and recreate it after the migration

    Incorrect

    Deleting removes the schedule and all its subscriber settings.

  • BRemove all subscribers from the schedule

    Incorrect

    The dashboard would still update, and the subscriber list would be lost.

  • CUnpublish the dashboard

    Incorrect

    The schedule menu offers a direct way to stop runs while keeping settings.

  • DPause the schedule

    Correct

    Paused schedules keep their configuration and subscribers, and the dashboard does not update or send notifications until resumed.

From the dashboard's Schedule list you can edit, pause or delete a schedule. A paused schedule keeps its configuration and subscriber settings but does not update or notify until resumed; deleting removes the schedule and its subscribers.

Question 4 · choose 1

A Databricks SQL alert fired yesterday. Today its status shows OK. What does OK mean?

  1. AThe alert has never been triggered
  2. BAn error occurred while the alert query was being evaluated
  3. CThe latest run did not meet the condition and threshold
  4. DThe alert is paused until someone resumes it
Show the answer and why
  • AThe alert has never been triggered

    Incorrect

    OK describes only the latest run; the alert may have triggered before.

  • BAn error occurred while the alert query was being evaluated

    Incorrect

    That is the ERROR status.

  • CThe latest run did not meet the condition and threshold

    Correct

    OK means the latest execution did not meet the configured condition; it does not mean the alert was never triggered.

  • DThe alert is paused until someone resumes it

    Incorrect

    OK is an evaluation result, not a paused state.

Alert status is TRIGGERED, OK or ERROR. TRIGGERED means the latest execution met the condition and threshold; OK means it did not, which doesn't mean the alert was never triggered; ERROR means evaluation failed.

Databricks documentation

Question 5 · choose 1

On some runs an alert's query finds nothing, and the team wants those runs to end in a status of their own instead of OK. Which alert setting does that?

  1. AEmpty result state
  2. BNotify on OK
  3. CThe condition's threshold value
  4. DThe compute selector
Show the answer and why
  • AEmpty result state

    Correct

    Empty result state sets a special status to return when the query returns no results.

  • BNotify on OK

    Incorrect

    Notify on OK sends a notification when the alert returns to OK.

  • CThe condition's threshold value

    Incorrect

    The threshold is the value that triggers the alert; it does not set a status for a query that returns nothing.

  • DThe compute selector

    Incorrect

    Compute chooses the SQL warehouse that runs the alert query.

Advanced alert settings include Notify on OK and Empty result state, which sets a special status to return when the query returns no results.

Databricks documentation

Practice domain 6 →Practice all domains →