PMLE Automating and Orchestrating ML Pipelines Practice Question

Answer analysis

Option-by-option breakdown

For each option: why learners choose it and why it is or isn't the right answer here.

• ✗

  Raise an exception in the component code and catch it in the pipeline definition.

  Why it's wrong here

  Exceptions raised in the component code are not propagated to the pipeline orchestrator. The component runs in a container, and the only signal to the pipeline is the exit code. Catching exceptions in the pipeline definition is not possible because the pipeline definition is static and does not execute the component code directly.

• ✓

  Return a non-zero exit code from the component's container.

  Why this is correct

  In Vertex AI Pipelines, a component failure is indicated by a non-zero exit code from the container. This causes the pipeline task to fail, and by default, the pipeline stops executing subsequent tasks. This is the standard way to enforce validation gates and prevent downstream tasks from running with invalid data.

• ✗

  Write a warning to the logs and continue execution.

  Why it's wrong here

  Writing a warning does not stop the pipeline. The pipeline would continue to the training step, potentially training on invalid data. The requirement is to fail the pipeline immediately, so logging a warning is insufficient. The component must signal failure through its exit status.

• ✗

  Use a conditional branch in the pipeline to skip training if validation fails.

  Why it's wrong here

  A conditional branch can skip training based on a condition, but it requires the validation component to output a boolean or a metric that the condition evaluates. This adds complexity and does not immediately fail the pipeline. The scenario asks for the pipeline to stop immediately, which is more directly achieved by failing the component.

Go deeper

Related to this question