Test severity levels in dbt - Practice Problems & Coding Challenges

Challenge - 5 Problems

🧠 Conceptual

intermediate

Understanding dbt test severity levels

In dbt, tests can have different severity levels that affect how failures are handled during runs. Which severity level causes dbt to fail the entire run if the test fails?

A pass

B error

C info

D warn

💡 Hint

Think about which severity level means the test failure is critical enough to stop the process.

The error severity level causes dbt to fail the entire run if the test fails. The warn level only logs a warning but continues the run.

❓ Predict Output

intermediate

Output of dbt test with warn severity

Given a dbt test configured with severity: warn and the test fails, what is the expected behavior in the dbt run output?

dbt

version: 2
models:
  - name: customers
    tests:
      - unique:
          column_name: id
          config:
            severity: warn

A The run logs a warning but continues to completion.

B The run fails and stops immediately.

C The run ignores the test and does not log anything.

D The run logs an error and retries the test.

💡 Hint

Severity 'warn' means the test failure is noted but does not stop the run.

When a test has severity 'warn' and fails, dbt logs a warning message but continues running other models and tests.

❓ data_output

advanced

Number of failed tests by severity

After running dbt tests, you get a summary table with test results and their severity levels. How many tests with severity 'error' failed if the table below is given?

test_name       | severity | status
unique_id       | error    | fail
not_null_email  | warn     | pass
unique_email    | error    | fail
accepted_values | warn     | fail

A 2

B 3

C 1

D 0

💡 Hint

Count only tests with severity 'error' and status 'fail'.

There are two tests with severity 'error' that failed: unique_id and unique_email.

🔧 Debug

advanced

Identify severity misconfiguration causing unexpected run failure

A dbt test is configured as follows but the run fails unexpectedly when the test fails. Identify the misconfiguration.

tests:
  - not_null:
      column_name: user_id
      severity: warn

However, the run fails on this test failure.

A Severity must be uppercase: <code>WARN</code> instead of <code>warn</code>.

B Severity 'warn' is not supported and causes failure.

C The test name is incorrect; it should be <code>not_null_test</code>.

D Severity should be set under <code>config</code> block, not directly under test.

💡 Hint

Check where severity is placed in the test configuration.

Severity must be set inside a config block for the test, not directly under the test name. Placing it incorrectly causes dbt to ignore it and default to 'error'.

🚀 Application

expert

Choosing severity levels for different test scenarios

You manage a dbt project with critical and non-critical data quality tests. Which severity level should you assign to tests that must never fail without stopping the pipeline, and which to tests that should only log issues but allow the pipeline to continue?

A Critical tests: warn; Non-critical tests: error

B Critical tests: pass; Non-critical tests: warn

C Critical tests: error; Non-critical tests: warn

D Critical tests: info; Non-critical tests: error

💡 Hint

Think about which severity stops the pipeline and which only logs warnings.

Critical tests should have severity 'error' to stop the pipeline on failure. Non-critical tests should have severity 'warn' to log issues but continue.