Some pipeline failures occur during validation or rendering for example, invalid YAML or unresolved expressions before execution starts. These failures do not generate console logs, so they cannot be captured or sent to external systems like Splunk. We need a way to expose these errors in logs or via API so they can be monitored and ingested like runtime failures.