Monitoring with Netdata

Netdata provides real-time system metrics (CPU, RAM, disk, network, services) with a lightweight web dashboard. This guide covers installation via script on a HolyCloud Linux VPS and hardening the firewall on port 19999.

Prerequisites

• Linux VPS with root SSH access
• At least 512 MB free RAM for the agent
• Ports: 19999 (local web interface or restricted)

Installation via the official script

Update the system, then run the one-liner installer (check the official Netdata page for the latest command):

apt update && apt install -y curl
curl https://get.netdata.cloud/kickstart.sh -o /tmp/netdata-kickstart.sh
sh /tmp/netdata-kickstart.sh --non-interactive

Variant with common options (no cloud claim if you do not use Netdata Cloud):

sh /tmp/netdata-kickstart.sh --non-interactive --stable-channel

Verify the service:

systemctl enable --now netdata
systemctl status netdata

Local dashboard access

By default, Netdata listens on all interfaces:

ss -tlnp | grep 19999
curl -s http://127.0.0.1:19999/api/v1/info | head

Test from your workstation via SSH tunnel (recommended before opening the port):

ssh -L 19999:127.0.0.1:19999 root@VOTRE_IP_VPS

Then open http://localhost:19999 in the browser.

Restrict listening (optional)

To listen only locally, edit /etc/netdata/netdata.conf ([web] section):

[web]
    bind to = 127.0.0.1

Restart:

systemctl restart netdata

Firewall: allow port 19999 carefully

UFW (Ubuntu/Debian)

Allow the port only from a fixed admin IP:

ufw allow from VOTRE_IP_ADMIN to any port 19999 proto tcp comment 'Netdata'
ufw deny 19999/tcp
ufw reload
ufw status numbered

firewalld (if used)

firewall-cmd --permanent --add-rich-rule='rule family="ipv4" source address="VOTRE_IP_ADMIN" port port="19999" protocol="tcp" accept'
firewall-cmd --reload

Security: do not expose 19999 publicly without authentication. Prefer SSH tunnel or a reverse proxy with authentication.

Metrics and retention

Check configuration:

grep -E 'memory mode|history' /etc/netdata/netdata.conf

Useful logs:

journalctl -u netdata -f --no-pager

Uninstall

/opt/netdata/bin/uninstall-netdata.sh --yes
# or per the method indicated by the install script

Troubleshooting

| Symptom | Action |

|----------|--------|

| Page unreachable | systemctl restart netdata, check bind to and UFW |

| High CPU | Reduce collection in plugins or increase VPS resources |

| Disk space | Adjust retention in netdata.conf |

Need help?

For centralized monitoring (Prometheus, Grafana), contact HolyCloud support for best practices on your plan.