Node.js SaaS teams should split error tracking and structured logs in catalog pipelines

The article argues that exception tracking and structured logging serve different purposes in nightly catalog import pipelines. Error trackers group recurring exceptions so engineers can prioritize fixes, while structured logs record the business steps of a run, such as rows read and rejected. Both should share a trace_id, but neither should be asked to replace the other. Structured logs also catch non-crash failures, like a completed import that yields zero active SKUs. The author describes a minimal implementation that writes JSON Lines to stdout and sends the same events to a logging endpoint, with retry logic and idempotency keys.