[Story 3/3] Tutorial: High-Throughput IoT Telemetry Pipeline with AMQP Workers & AI Diagnostics

[crapougnax]

🎯 Story Overview

As a Distributed Systems & IoT/AI Architect,
I want to build an industrial telemetry acquisition pipeline that decouples sensor frame ingestion via RabbitMQ workers, manages physical device capabilities via MDM ontology, routes data across polyglot backends (PostgreSQL, AMQP, S3), and triggers structured Gemini AI incident diagnoses on threshold breaches,
So that high-frequency edge IoT workloads run asynchronously, resiliently, and sovereignly on on-premise or cloud infrastructure.

---

📦 Ecosystem Packages

• @quatrain/mdm (Device Ontology, Hardware Capabilities Matrix)
• @quatrain/queue & @quatrain/queue-amqp (RabbitMQ Broker, Worker Concurrency)
• @quatrain/backend & @quatrain/backend-postgres (Device Registry & Telemetry Relational Store)
• @quatrain/storage & @quatrain/storage-s3 (Cold Storage Telemetry Archiving)
• @quatrain/ai & @quatrain/ai-gemini (Google Gemini, Schema-Guaranteed Diagnostic JSON)
• @quatrain/auth-rbac (TarpitManager Anti-Abuse Rate Limiting)

---

📋 Child Tasks & Implementation Plan

• feat(example-iot-telemetry): define MDM device ontology and telemetry metric models #42: feat(example-iot-telemetry): define MDM device ontology and telemetry metric models
• feat(example-iot-telemetry): configure AMQP RabbitMQ broker with concurrent worker consumers #43: feat(example-iot-telemetry): configure AMQP RabbitMQ broker with concurrent worker consumers
• feat(example-iot-telemetry): implement polyglot ObjectUri pipeline and Gemini AI anomaly synthesis #44: feat(example-iot-telemetry): implement polyglot ObjectUri pipeline and Gemini AI anomaly synthesis
• docs(example-iot-telemetry): write comprehensive step-by-step tutorial, Containerfile and compose stack #45: docs(example-iot-telemetry): write comprehensive step-by-step tutorial, Containerfile and compose stack

---

✅ Acceptance Criteria

1. MDM model supports multi-axis hardware capabilities (radio, power, sensor bus).
2. RabbitMQ workers ingest sensor frames asynchronously with configurable concurrency (concurrency: 10).
3. Endpoint protected against sensor flooding via TarpitManager.
4. Anomaly detection invokes Gemini AI via generateStructured to output guaranteed diagnostic JSON schemas.
5. Multi-arch Containerfile and docker-compose.yml for local stack (PostgreSQL + RabbitMQ + MinIO).
6. Full runnable reference application under examples/iot-telemetry-engine.