andreas404/insightd-hub repository overview

⁠Insightd Hub

Server awareness without the overhead. Central hub for insightd⁠ — a self-hosted monitoring tool for homelabbers.

Monitors Docker containers, hosts, and HTTP endpoints across multiple servers. Includes a modern React dashboard, smart alerts, weekly digest emails, and webhook notifications.

⁠Quick Start

⁠Standalone (single host, no MQTT needed)

docker run -d \
  --name insightd \
  --restart unless-stopped \
  -p 3000:3000 \
  -v /var/run/docker.sock:/var/run/docker.sock:ro \
  -v /:/host:ro \
  -v insightd-data:/data \
  -e TZ=UTC \
  andreas404/insightd-hub:latest

Open http://localhost:3000 and follow the setup wizard.

⁠Multi-host (with MQTT + agents)

# docker-compose.yml
services:
  mosquitto:
    image: eclipse-mosquitto:2
    restart: unless-stopped
    ports:
      - "1883:1883"

  hub:
    image: andreas404/insightd-hub:latest
    restart: unless-stopped
    depends_on:
      - mosquitto
    ports:
      - "3000:3000"
    volumes:
      - hub-data:/data
    environment:
      INSIGHTD_MQTT_URL: mqtt://mosquitto:1883
      INSIGHTD_ADMIN_PASSWORD: changeme
      TZ: UTC

volumes:
  hub-data:

Then deploy andreas404/insightd-agent⁠ on each remote host.

⁠Features

• Dashboard — health scores, availability, unified "Needs Attention" feed
• Multi-runtime — Docker and Kubernetes/k3s (via the agent's DaemonSet mode)
• Host grouping — collapsible Hosts page sections by cluster, environment, or location
• Container monitoring — status, CPU, RAM, restarts, network/block I/O, health checks
• Host metrics — CPU, memory, load, disk, GPU, temperature (Docker mode); CPU, memory, uptime from kubelet (k8s mode)
• HTTP endpoint monitoring — uptime tracking, response times
• Insights engine — time-of-day baselines, anomaly detection, predictive alerts, correlations
• Real-time alerts — 10 alert types with cooldowns and auto-resolution
• Webhooks — Slack, Discord, Telegram, ntfy, or generic
• Weekly digest emails — HTML + plaintext summaries
• Container actions — start/stop/restart from the UI (opt-in, Docker mode)
• Stacks — container groups across hosts (auto-detected from Docker Compose, or create your own)
• Public status page — shareable uptime page at /status (opt-in)
• API keys — programmatic access with hashed storage

⁠Configuration

All via environment variables. Key ones:

See .env.example⁠ for the full list (40+ variables).

⁠Volumes

⁠Architecture

[Agents] --> MQTT (Mosquitto) --> [Hub] --> SQLite --> React UI + Alerts + Digests

• Hub mode: connects to MQTT, receives metrics from agents
• Standalone mode: no MQTT, collects locally (single host)

⁠Resources

• ~28MB RAM typical usage
• SQLite storage, no external database
• Multi-arch: linux/amd64, linux/arm64

⁠Links

• GitHub⁠
• Agent image⁠
• Security policy⁠