Challenge
Monitoring vital signs and fatigue risk for field engineers in physically demanding environments requires ingesting continuous wearable telemetry bursts while pushing sub-second threshold alerts to supervisor dashboards.
Amazfit: Polyglot Wearable Telemetry Platform
High-throughput field-engineer health monitoring combining a C++ userver ingestion tier, NestJS WebSocket fan-out, and a React 19 PWA.
- C++ userver
- NestJS & Prisma
- React 19 PWA
- Redis Pub/Sub & MQTT
- Wearable IoT
Confidential
Approach
We architected a polyglot three-service topology: a C++ userver ingestion service (built reproducibly with Nix) handles machine-facing watch payloads, computes health scores, persists time-series records to PostgreSQL, and publishes threshold alerts to Redis; a NestJS gateway fans alerts out over Socket.io rooms to an offline-capable React 19 + Tailwind v4 PWA.
Outcome
Separated high-frequency machine ingestion from human dashboard workflows, enabling deterministic alert evaluation and multi-tenant team risk visibility.
Results
- Polyglot C++ userver ingestion tier paired with NestJS and React 19 PWA
- Real-time Redis Pub/Sub to Socket.io alert fan-out for field supervisor rooms
- Reproducible Nix + Docker Compose build and single-VM deployment pipeline