ETL Pipeline Error Handler

Pipeline {{pipeline_name}} failed on {{failure_date}} at stage {{failure_stage}}. Error: {{error_message}}

For this failure:
1. Error classification — is it a transient issue (retry), data issue (validation), or infrastructure problem (scaling)?
2. Impact scope — which downstream datasets are stale? For how long?
3. Data consistency — can we safely rerun from this point, or restart from the beginning?
4. Recovery path — should we use cached intermediate data, backfill, or wait for next run?
5. Prevention — add monitoring, add data validation, or increase resource allocation?
6. Owner notification — which teams depend on {{affected_table}}?

Provide step-by-step recovery runbook.