Testing and debugging

3 h
requires JVM-01
requires JVM-02

From your world. dbt tests + a Spark pipeline integration test, but as JUnit/AssertJ and reading JVM stack traces

Prompt A. "Explain to a dbt/pytest user how you would test a Kafka producer/consumer in Java. Say which part is a fast JUnit/AssertJ unit test. Say which part needs a real broker. Say how you grade it without Docker. Say why you poll instead of sleep."

Prompt B. "You have a stack trace with several Caused by: sections. Walk through how to find the actual bug. Say which line is the root cause. Say which frame you open first. Say why reading top-down is a mistake."

Rubric — a strong explanation hits these (5–7):

Uses JUnit 5 @Test + AssertJ fluent assertions and the float within tolerance.
Uses assertThatThrownBy to assert exceptions.
Distinguishes unit tests (pure logic) from integration tests (real broker).
States the no-Docker rule: the native broker or TopologyTestDriver grades the exercise. Testcontainers is an extra that skips when Docker is down.
Polls for expected records with a deadline instead of sleep().
Identifies the deepest Caused by: as the root cause. Identifies the first app-package frame as the place to look. Notes that ... N more leaves out frames.
(Bonus) Mentions Java 21 helpful NPE messages.