Update existing extractors to emit richer relationships

[ravisuhag]

Context

Many existing extractors already have access to relationship data beyond lineage but don't surface it. Once the typed relationship model is in place, extractors should emit everything they can.

Scope

Audit and update existing extractors to capture available relationships:

• BigQuery — table read/write patterns from access logs, dataset ownership
• Tableau — dashboard-to-datasource relationships, workbook ownership
• Optimus — job ownership, job-to-table read/write patterns
• Kafka — topic ownership, consumer group relationships
• Postgres/MySQL — view dependencies, role-based access patterns
• Grafana/Metabase/Superset — dashboard ownership, datasource connections

Approach

For each extractor:

1. Identify what relationship data is available at the source
2. Map source relationships to typed edges from the new relationship model
3. Emit relationships as raw observations alongside existing asset data — Compass handles entity resolution
4. Update tests and documentation

Why

The richest graph comes from extracting relationships that already exist at the source. This is low-hanging fruit — the data is there, we just aren't capturing it.

Dependencies

• Typed relationship model (Expand relationship model with typed edges beyond lineage #499)
• Proto schema changes in raystack/proton

References

• Meteor Roadmap — From Records to Graph