Real-Time Spectral Data Processing Pipeline in Node.js

Design a high-performance Node.js microservice for processing large-scale spectroscopic data streams from scientific instruments. Create a streaming architecture that can handle raw spectral data from multiple sensor inputs, perform real-time frequency domain transformations using complex mathematical libraries, and generate normalized JSON output. Implement error handling for instrument signal drift and include performance monitoring for data ingestion rates. The solution must support concurrent processing of multi-channel spectral data with less than 50ms latency.