Performance Benchmarks #39

Problem

As we optimize the batch processor and NATS handling, we have no way to prove that changes actually improve speed or to catch "performance regressions" where a new feature accidentally slows down ingestion by 50%.

Proposed Solution

Add Go benchmarks (go test -bench) for critical paths: batch serialization, NATS message ingestion, and SQL template rendering. These will provide a baseline to ensure WaveHouse remains high-performance as it grows.

Alternatives Considered

Ad hoc testing with high volume data, but benchmarks provide a more stable and repeatable metric.

Additional Context

None