Your question is Debugging Distributed Pipeline Bottlenecks. Take a moment with it on the right.
Talk me through your thinking if you like. When you're confident, submit your answer and I'll grade it like a real screen (7/10 or better passes).
How would you debug a performance bottleneck in a distributed data pipeline?
Explain how you would isolate the slow stage, distinguish infrastructure limits from data-related issues, and validate the fix. Cover the metrics, logs, query profiles, and tracing information you would inspect across ingestion, processing, Snowflake loading, and orchestration. Discuss how you would handle common causes such as data skew, insufficient parallelism, inefficient transformations, backpressure, warehouse queuing, and storage layout problems.