Real-time Telemetry Data Pipeline

Design an LLD for a telemetry processor that ingests streams of data from IoT sensors. The system must: 1. Cleanse data (remove nulls), 2. Normalize units (Celsius to Fahrenheit), and 3. Route data to multiple sinks (a Time-Series Database, a Real-time Dashboard, and an Alerting Engine). The processing steps should be configurable per sensor type, and new processing stages (e.g., Anomaly Detection) should be added easily.