Access All Ingested Data in ES|QL for Flexible Analysis (GA)

## Value Proposition

Break free from rigid schema constraints and access every piece of data you've ingested — mapped or not — directly within ES|QL queries. This eliminates the "ignorance cliff" by ensuring that all fields are queryable, providing flexibility for ad-hoc exploration, threat hunting, and troubleshooting across diverse and evolving datasets without costly reindexing.

## Expected Outcome

Access to unmapped fields in ES|QL shipped in Tech Preview in Elasticsearch 9.4, with two modes: NULLIFY (return nulls for missing fields) and LOAD (fetch values from _source at query time).

For GA, we are hardening both modes for production use:

- **Full-text search compatibility:** LOAD mode works alongside MATCH and other full-text functions on mapped fields
- **Branching command support:** LOAD works correctly with subqueries, views, and FORK
- **Performance validation:** nightly benchmarks and heap-safety testing to ensure LOAD scales gracefully and circuit-breaks correctly under adversarial workloads
- **Backward compatibility:** version-gated autocasting behavior for safe mixed-version cluster upgrades
- **Correctness:** resolved interactions with LOOKUP JOIN, partially unmapped fields, and type conflicts across indices