Reviewed by Aditya Kumar · Last reviewed 2026-03-24
Architectural logic: Retries handle transient failures; timeouts prevent hung tasks from blocking DAGs. Configuration: default_args or per-task: retries, retry_delay, retry_exponential_backoff, execution_timeout. Example: retries=3, retry_delay=timedelta(minutes=2),...
Architectural logic: Retries handle transient failures; timeouts prevent hung tasks from blocking DAGs. Configuration: default_args or per-task: retries, retry_delay, retry_exponential_backoff, execution_timeout. Example: retries=3, retry_delay=timedelta(minutes=2), retry_exponential_backoff=True, execution_timeout=timedelta(minutes=45). Why exponential backoff: Reduces thundering herd on downstream services. Scalability: Too many retries on large DAGs can delay subsequent runs. Cost: Failed tasks still consume worker time. Best practice: Use exponential backoff; set timeouts; on_failure_callback for alerts; trigger_rule for downstream behavior.
Red Flag: No retries or infinite timeout—production anti-pattern. Pro-Move: 'We use exponential backoff and 45-min timeout; failures trigger PagerDuty via callback—reduced false alerts by 60%.'
According to DataEngPrep.tech, this is one of the most frequently asked Cloud/Tools interview questions, reported at 1 company. DataEngPrep.tech maintains an editor-reviewed database of 1,863 data engineering interview questions across 7 categories.