Databricks Production Checklist

Infrastructure

• Production cluster pool configured
• Instance types validated for workload
• Autoscaling configured appropriately
• Spot instance ratio set (cost vs reliability)

Step 2: Code Quality Verification

Testing

• Unit tests passing
• Integration tests on staging data
• Data quality tests defined
• Performance benchmarks met

# Run tests via Asset Bundles
databricks bundle validate -t prod
databricks bundle run -t staging test-job

# Verify test results
databricks runs get --run-id $RUN_ID | jq '.state.result_state'

Code Review

• No hardcoded credentials
• Error handling covers all failure modes
• Logging is production-appropriate
• Delta Lake best practices followed
• No collect() on large datasets

Step 3: Job Configuration

# resources/prod_job.yml