How do you optimize a Lucene index for search performance?

Search performance tuning in Lucene usually comes down to a handful of high-leverage decisions rather than one silver bullet:

1. Choose field types deliberately - use DocValues fields for sorting/faceting instead of loading stored values at query time.
2. Disable norms/term vectors on fields where scoring precision or highlighting isn't needed, to save memory.
3. Use Points fields (IntPoint, LongPoint) for numeric range filters rather than string-based range queries.
4. Tune the MergePolicy so segment count stays reasonable without excessive merge I/O.
5. Warm searchers after reopening so caches (like DocValues) aren't cold on the first real query.
6. Favor filters over scored queries for non-relevance criteria, since filters can be cached and skip scoring work.

Profiling with actual production-shaped queries before and after each change matters more than applying every tuning option blindly - some of these trade indexing speed or memory for query speed, so the right mix depends on the workload.