Netdata 2.10.1

Netdata collects hundreds of metrics per second and shows them in a dashboard, with no configuration required.

curl https://get.netdata.cloud/kickstart.sh > /tmp/netdata-kickstart.sh
sh /tmp/netdata-kickstart.sh

Open port 19999 and it is all there: CPU per core, memory breakdown, disk I/O per device, network per interface, plus automatic detection of nginx, PostgreSQL, Docker containers, and dozens of others.

Per-second resolution

The distinguishing feature. Most monitoring scrapes every 15 or 30 seconds, which averages away brief spikes entirely.

A process that saturates a disk for two seconds is invisible at a 30-second interval and obvious at one second. For diagnosing “the machine froze for a moment” this is the difference between an answer and a shrug.

The tradeoff against Prometheus

Netdata answers “what is happening right now” better than anything else and keeps a limited window by default.

Prometheus with Grafana answers “what happened last Tuesday” and gives you alerting rules across many machines.

Running both is common and reasonable: Netdata for detailed investigation on a specific host, Prometheus for the long view. Our monitoring guide covers the split.

Netdata can also export metrics to Prometheus, so the two compose rather than compete.

Do not expose it

The dashboard has no authentication by default and reveals a great deal about the machine.

Bind it to localhost and reach it through a reverse proxy with authentication, or over a VPN:

# /etc/netdata/netdata.conf
[web]
    bind to = 127.0.0.1

Alerting

Netdata ships with health checks already configured and can notify by email, Slack, ntfy, Discord, and others through health_alarm_notify.conf.

Configure at least one notification path. A dashboard nobody looks at is not monitoring.

Verify Your Download

Release archives on GitHub are published with checksums and signatures.