Introduction
Komari is a lightweight, self-hosted server monitoring tool that provides real-time visibility into your infrastructure with minimal resource overhead. Its clean web dashboard shows CPU, memory, disk, and network metrics across all your servers without complex configuration.
What Komari Does
- Monitors CPU, memory, disk, and network on Linux and macOS hosts
- Displays real-time metrics on a responsive web dashboard
- Sends alerts via webhook, Telegram, or email on threshold breaches
- Tracks historical data for trend analysis
- Provides remote terminal access for quick troubleshooting
Architecture Overview
Komari uses a lightweight Go agent on monitored hosts that reports metrics to a central server via WebSocket. The server stores time-series data in an embedded database and serves the web dashboard as a single-page application. The agent binary is under 10MB and uses minimal CPU and memory.
Self-Hosting & Configuration
- Deploy server via Docker or single binary on any Linux host
- Agent installs with a one-line script on monitored machines
- Configure alert thresholds in the web UI or config file
- Supports reverse proxy with Nginx or Caddy for HTTPS
- Data retention configurable from days to months
Key Features
- Sub-10MB agent with negligible resource consumption
- One-command agent installation with auto-registration
- Mobile-responsive dashboard for on-the-go monitoring
- Multi-server overview with sorting and grouping
- No external database dependency (embedded storage)
Comparison with Similar Tools
- Beszel — similar lightweight approach; Komari adds remote terminal access
- Netdata — feature-rich but heavier; Komari focuses on simplicity
- Uptime Kuma — primarily uptime checks; Komari provides full system metrics
- Grafana + Prometheus — powerful but complex setup; Komari works out of the box
FAQ
Q: What operating systems does the agent support? A: Linux (x86_64, ARM64) and macOS. Windows support is experimental.
Q: How much disk space does historical data use? A: Approximately 1MB per host per day at default collection intervals.
Q: Can I monitor Docker containers? A: Yes, container-level metrics are collected when Docker is detected.
Q: Is there an API for custom integrations? A: Yes, a REST API provides all metrics and alert management programmatically.