Netdata: what it is
Netdata is an open-source monitoring agent that collects metrics every second and shows them on a local web dashboard. You install the agent on a server, desktop or container host, and it starts discovering what is running: CPU, memory, disks, network interfaces, processes, containers and many common services. The project describes per-second collection, built-in alerts and anomaly detection with machine-learning models trained on each node.
It suits admins who want detailed live data from a few machines quickly, without first building a Prometheus and Grafana stack. For long-term, fleet-wide storage you can still stream or export Netdata metrics to other systems. We have not run our own benchmarks of Netdata (Not tested yet); this page summarises the official documentation.
Netdata at a glance
| Item | Details |
| Developer | Netdata Inc. and the open-source community (GitHub: netdata/netdata) |
| Latest stable version | v2.12.2, released on GitHub on 23 October 2026 |
| Platforms | Linux, FreeBSD, macOS and Windows (64-bit). Linux has the widest coverage; some collectors, such as systemd services, containers and live sockets, are Linux-only |
| Licence | Netdata Agent: GPL v3 or later. The dashboard UI is free to use but ships under Netdata’s own NCUL1 licence, not an open-source one |
| Dashboard | http://localhost:19999 (or http://NODE:19999 from another machine) |
| Windows package | netdata-x64.msi, about 176 MB for v2.12.2 |
| Last verified | 27 October 2026, against the Netdata GitHub repository and documentation |
What Netdata monitors
- System resources: CPU, memory, load and shared resources on all four supported operating systems.
- Storage and network: disks, mount points, filesystems, RAID arrays, network interfaces and protocols.
- Processes: resource use per process, including out-of-memory events on Linux.
- Logs: the systemd journal on Linux, and Windows Event Log and ETW on Windows.
- Containers and VMs: Docker, containerd, LXC/LXD and Kubernetes on Linux; virtual machines from the host via cgroups on Linux and Hyper-V on Windows.
- Applications: packaged collectors for nginx, Apache, PostgreSQL, Redis, MongoDB and hundreds more, plus OpenMetrics and StatsD for your own applications.
- Hardware: on Linux, sensors such as fans, temperatures, GPUs, IPMI, S.M.A.R.T. and NVMe.
How to install Netdata
Linux
The documented method is the one-line kickstart.sh installer from get.netdata.cloud, which works on most distributions and lets you choose the stable or nightly release channel with --release-channel. The documentation also publishes the script’s checksum so you can verify it before running it. Docker and Kubernetes installs are documented separately.
Windows
Download netdata-x64.msi from the latest GitHub release and run it with administrator rights. Netdata recommends Windows 10 or 11, or Windows Server 2019 or newer, as the supported baseline. The installer may start on older 64-bit Windows, but those versions are not tested or officially supported.
macOS and FreeBSD
Both have their own installation pages in the Netdata documentation. Fewer collectors are available there than on Linux.
After installation, open http://localhost:19999 in a browser. The agent discovers most metrics automatically; individual collectors can be configured afterwards.
Free use and Netdata Cloud
The agent and its local dashboard work without an account. Netdata Cloud adds sign-in, central views and paid plans (Homelab, Business, Enterprise On-Premise). The access-control page in the documentation lists what changes between levels: real-time metrics, history, charts, anomaly detection and alert notifications are available at every level, while node dashboard access is listed as 5 nodes for anonymous and free Community users and unlimited on paid plans. Some sensitive live functions, such as process details and network connections, require signing in. Check that page before planning a larger rollout, because these rules can change.
Netdata vs Zabbix: which to start with
Netdata focuses on high-resolution live data from each node with very little setup. Zabbix is a central server with a database, templates, discovery and long-term history for many hosts and network devices. Many teams use both: Netdata for live troubleshooting on a node, and a central system for fleet-wide history and reporting. For tracing distributed applications, look at SigNoz instead.
Limitations to know
- The UI licence (NCUL1) is not open source, even though the agent is GPL v3+.
- Collector coverage on Windows, macOS and FreeBSD is smaller than on Linux.
- Free use has documented limits on node dashboard access and on some sensitive functions.
- Per-second collection on many metrics uses some CPU, memory and disk on each node. Disable collectors you do not need on small machines.
FAQ
Is Netdata free?
The agent is free and open source under GPL v3 or later, and the local dashboard is free to use under the NCUL1 licence. Netdata Cloud has a free Community tier and paid plans.
Does Netdata run on Windows?
Yes. There is a 64-bit MSI installer, and Windows support includes system resources, storage, network, processes, Windows Event Log and Hyper-V metrics.
What port does Netdata use?
The dashboard and API listen on port 19999 by default.
What is the latest Netdata version?
On 27 October 2026 the latest stable release on GitHub was v2.12.2 (23 October 2026). Nightly builds are published separately.
Can Netdata replace Prometheus and Grafana?
For live monitoring of a handful of nodes it often can, because it collects, stores, alerts and visualises on its own. For long-term, multi-team storage and custom dashboards, many teams keep a central system and export Netdata metrics to it.
Sources
- Netdata releases (GitHub)
- Netdata README: platforms, licences, getting started (GitHub)
- Install Netdata on Windows (Netdata documentation)






