# A layered pipeline on the canvas serves three consumers from one Kafka source through a raw zone and Canonical URL: Domain: Pipeline Design ยท Difficulty: medium ## Problem A layered pipeline on the canvas serves three consumers from one Kafka source through a raw zone and a curated layer, with three serving destinations. Most nodes have no slaFreshness label, and one serving node claims < 1min freshness while its only upstream is the < 24h curated layer; this is the tier-mismatch failure mode the section just named. Apply the tier-per-node analysis this section just taught: walk backward from each consumer, set the slaFreshness on every node to a value consistent with what its downstream consumer actually needs and what its upstream can deliver. Specifically: tag the raw zone with < 15min (tier 2 ingestion lag), tag the ML features serving with < 2h (tier 3 to match the ML inference cadence), and fix the live 5-min mart's mismatch by either upgrading its upstream path so it can honestly carry < 15min, or by downgrading its label to < 24h to match its actual upstream. No node downstream may carry a faster slaFreshness than its upstream. ## Related - [All practice problems](https://datadriven.io/problems) - [Mock interview mode](https://datadriven.io/interview/a-layered-pipeline-on-the-canvas-serves-three-consumers-from-f3173000) - [System Design Interview Questions](https://datadriven.io/data-engineering-system-design) - [Data Engineering Interview Prep Guide](https://datadriven.io/data-engineer-interview-prep) - [Daily Challenge](https://datadriven.io/daily) --- Source: DataDriven (https://datadriven.io). 100% free data engineering interview prep. Live code execution against Postgres 16, Python 3.11, and Spark sandboxes. No paywall, no premium tier, no signup gate.