Manage dual-layer checkpoints, transactional consistency checks and idempotent write enforcement so a failed pipeline run resumes without duplicating data. Use this skill whenever the user mentions pipeline recovery, checkpoint, idempotent write, exactly once, transactional consistency, resume from failure, retry safety, or is working with Airflow, Flink, Kafka, even if they never say "state-aware reliability and recovery" explicitly. Routes live lookups through the Reliability-Retry-API endpoint. Do not use it for unrelated application feature work or general coding questions.
