6d1d9c8620
- telegraf.conf: fix Mosquitto/BunkerM container names (hyphens not underscores) - tegraf.conf: comment out BunkerM consumer (auth fails, simulator not sending) - openremote/docker-compose.yml: switch PG image to timescaledb-ha:pg15 (fixes timescaledb_toolkit crash) - Add session_resume + architecture snapshot 2026-05-25 - Update TODO.md with current status
3.3 KiB
3.3 KiB
Smart City Digital Twin — Architecture État Actuel (snapshot)
Date: 2026-05-25 ~18:45 UTC Total containers: 117
📡 Pipeline Données (IOT → Stockage → Visualisation)
┌──────────────┐ MQTT topics: ┌──────────┐ ┌───────────┐ ┌────────────┐
│ Simulator │──── smartcity/●/● ──▶│ EMQX │────▶│ Telegraf ────▶│ InfluxDB │
│ (60 sensors)│ " │ (v5.0) │ │ (mqtt_ │ │ (v2.7) │
│ │──── smartcity/●/● ──▶│ │ │ consumer)│ │ │
│ ENABLE_BUNKER│ " │ │ │ influxdb │────▶│ bucket: │
│ =0 (off) │ │ │ │ _v2) │ │ "smartcity"│
└──────────────┘ └──────────┘ └───────────┘ └─────┬──────┘
│
Mosquitto (chirpstack) ◀── allow_anonymous=true ── config mosquitto.conf │
(no data visible in logs yet) │
▼
┌────────────┐
│ Grafana │
│ Dashboard │
│ Smart City │
└────────────┘
🏗️ Stacks actives (principales)
| Stack | Containers | État | Réseau |
|---|---|---|---|
| OpenRemote | postgresql (timescaledb-ha), keycloak, manager | PG: starting, KC: healthy, Manager: starting | smartcity-shared + openremote_default |
| Smart City Core | simulator, telegraf, grafana, influxdb, prometheus | UP | smartcity-shared |
| FIWARE GIS | orion-ld, orionproxy | healthy | smartcity-shared |
| ChirpStack | chirpstack, mosquitto, postgres, redis | UP | dédié + smartcity-shared |
| Ditto | gateway, policies, mongodb | UP/restarting | dédié |
| Contexus | app (unhealthy), postgres (healthy), redis (healthy) | mixed | smartcity-shared |
| Mainflux | 15+ microservices, mqtt, jaeger (exited) | mixed | dédié |
| Stellio | api-gateway (UP), search+subscription+kafka+postgres (exited) | broken | dédié |
| Traefik | traefik:v3.1 | UP | traefik-public |
⚠️ Points connus
- OpenRemote : crash loop corrigé (image PG changée), à redémarrer
- Telegraf : fixé (noms containers + BunkerM désactivé)
- Stellio : cassé (kafka + postgres exited) — NGSI-LD lowest priority
- Contexus : app unhealthy
- Pulsar/Redpanda : arrêtés (stabilité CPU)
- BunkerM : unhealthy, port 1900 (port MQTT originel)
- Simulator : MQTT OK 1/2 — un seul broker sur deux (Mosquitto?)