Snag Search Test — v3

Project: snag-search-test · RTDB path: /snagFeed ·

1a. Small seeds — single-shot writes

Single set() call. Safe at these sizes.

1b. Big seeds — batched writes (10k entries per chunk)

Streaming build + batched update(). Each chunk ~10k entries (~1.3 MB).

1c. Stress sizes — for finding the architecture's ceiling

Heads up: Seed times: 3M ≈ 4 min, 5M ≈ 6.5 min, 7M ≈ 9 min. Storage: 3M ≈ 384 MB, 5M ≈ 640 MB, 7M ≈ 896 MB (near 1 GB free-tier cap). Mobile Safari likely crashes during Load All at 3M+.

2. Load corpus into memory

No corpus loaded
Single get() on /snagFeed. At 7M entries this transfers ~900 MB and parses to ~1.8 GB JS heap.

3. Run searches (local filter, zero network)

4. Cleanup

Free up the test project's storage between size tests. Especially useful before/after the 5M+ runs.
Ready. Open browser DevTools console for parallel logs.