Discovery Pipeline Update: New Sources Now Prioritize Fresh Listings

What changed in the pipeline

The listing discovery workflow now does more than collect candidates. It evaluates them before publish so the queue is biased toward genuinely new resources.

The operational improvements

• Preflight requests identify create vs update before publish
• New listings are published before existing ones
• Source-level yield is logged so weak feeds are easier to spot
• Low-performing sources can be replaced faster

Why this matters

A discovery job that mostly updates existing records gives the illusion of activity without growing the catalog. Prioritizing creates gives the schedule a clearer job: expand the directory first, maintain second.

Resulting workflow

1. Discover candidate resources
2. Normalize and classify them
3. Preflight against the internal API
4. Publish new listings first
5. Review source stats and replace weak sources

This makes the ingestion loop more measurable and more practical to tune.