Who This Is For
Beginner.
Think of it like a dashcam versus a full flight recorder. Quick, immediate, great for "what's happening right now," not built for deep historical digging.
What You'll Build
One monitored server with at least one alert threshold, live within minutes.
Prerequisites
- A Linux host (Docker or bare metal)
When Netdata Is the Right Choice
When you want immediate, real-time troubleshooting visibility without the setup overhead of Prometheus + Grafana.
Install via Docker
$docker run -d --name netdata \
$ -p 19999:19999 \
$ -v netdataconfig:/etc/netdata \
$ -v netdatalib:/var/lib/netdata \
$ --cap-add SYS_PTRACE \
$ netdata/netdata
$ -p 19999:19999 \
$ -v netdataconfig:/etc/netdata \
$ -v netdatalib:/var/lib/netdata \
$ --cap-add SYS_PTRACE \
$ netdata/netdata
Dashboard Tour
- CPU
- Memory
- Disk
- Network
- Processes
- Containers (if monitoring the Docker socket)
Restrict Access
The dashboard shows detailed system internals, keep it off any public interface; local network or VPN only.
Create Health Alerts
Netdata ships with sane default alert templates for common thresholds; customize them under the health configuration directory.
Use It While Troubleshooting
- "Why is my service slow?", check CPU/IO in real time
- "Which process is using RAM?", the Processes view
- "Why is disk IO high?", the Disk panel, per-process breakdown
Security & Backup Notes
- Do not expose the Netdata web interface publicly, it reveals detailed internals about the host
Troubleshooting
- Docker socket access issues, container monitoring needs the socket mounted and the right capabilities
- Web interface exposed accidentally, double-check port bindings only allow local/LAN access
- Excessive alert noise, tune default thresholds down for a small homelab, they're tuned for larger systems by default
Lab Finish Line
One monitored server with at least one alert threshold.