Hnsw

How does an HNSW index handle deletes and updates under a heavy write workload?

Estimate the RAM cost of an HNSW index over 100M vectors at 1024 dimensions with M=32, including graph overhead.

As M grows in an HNSW index, four things move at once. Identify what climbs versus what slows.

Explain exactly what the ef_search knob on an HNSW index controls at query time.

Aggressive metadata filtering can silently destroy recall on an HNSW index even when matches exist. Why?