Add usage tracking to record entity query and access patterns

[ravisuhag]

Context

Compass doesn't track which entities are queried or accessed. Usage data is a strong signal for relevance — popular entities should rank higher, unused entities may be deprecation candidates.

Scope

• Record access patterns when entities are queried through Compass APIs
• Track: which entities are looked up, by whom, how frequently
• Expose usage metrics in entity responses and search ranking
• Aggregate usage data for trend analysis (growing/declining popularity)

Design Considerations

• Privacy: aggregate usage, don't expose individual access logs to other users
• Performance: access tracking should not add latency to the read path (async recording)
• Retention: configurable window for usage data (e.g., last 90 days)

Why

Usage data helps AI agents recommend the most relevant entities and identify authoritative sources. "This table is queried by 12 teams daily" is a stronger signal than "this table exists."

References

• Compass Roadmap — Active Knowledge