Throughput & Scale Audit

Verify that the system behaves correctly and remains responsive as work volume increases. This audit catches the class of bugs that only appear at real-world scale: slow queries, memory bloat, queue congestion, UI freezes, and background job starvation.

## Key Points

- 500+ assets (images, videos, documents, or domain-equivalent records)
- 25+ locations / categories / parent entities
- 50+ scenes / child entities per parent
- 20+ projects with varying sizes (1 asset to 200 assets)
- 5+ user accounts with different roles
- 10,000+ log/event records
- 100+ queued/completed jobs in history
1. Load the main listing page with 10 records. Count DB queries (via query log or ORM debug).
2. Load the same page with 500 records. Count DB queries.
3. Compare.
- [ ] Query count at 500 records is within 2x of query count at 10 records.
- [ ] Absolute query count per page load is under 20.

## Quick Example

```
10 records:  12 queries (1 list + 1 per record for author + 1 per record for thumbnail)
500 records: 1002 queries
--> FAIL: Classic N+1 on author and thumbnail relations
```

```
[ ] Separate queues for bulk vs interactive jobs
[ ] Worker concurrency limits per job type
[ ] Priority levels implemented (at least: critical, normal, bulk)
[ ] Dead letter queue for repeatedly failing jobs
[ ] Backpressure mechanism (reject new jobs when queue > threshold)
```