Netdata is a distributed, real-time observability platform designed to collect, store, and analyze system metrics with per-second granularity and sub-two-second latency.[4] Unlike centralized monitoring solutions, Netdata processes data at the edge using lightweight agents that include machine learning-based anomaly detection trained locally on each monitored node.
Netdata Agent, a daemon responsible for collecting and storing observability data, running machine learning for anomaly detection, querying observability data and evaluating alerts, in real-time.
Netdata Parent, for creating observability centralization points (uses the same software as Netdata Agents).
Netdata Cloud, a commercial control plane offering enterprise features.
Netdata UI, a web application responsible for all dashboards, using as backend Netdata Agents, Parents or Cloud.
It is mostly written in C, Go, Rust, Python and JavaScript, and aims to use minimal system resources.
Netdata is designed to be installed on a system without interrupting any of the applications running on it.
Netdata can be configured to store its time-series data either in memory or on disk. By default, it uses its database engine (dbengine) to store data on disk with tiered retention at multiple resolutions; it can also be configured for in-memory-only operation. [5][6]
By default it contains plugins that collect system, container and application metrics. Its behavior is extensible by using its plugin API. It also provides plugins that scrape Prometheus/OpenMetrics endpoints and ingest OpenTelemetry metrics and logs.
Dashboards are automatically generated by Netdata. It has an interface with customizable themes.
There are no dependencies, as it operates as its own web server, with static web files.
Netdata is known for its distributed design.[7] Instead of funneling all data into a few central databases like most traditional monitoring solutions, Netdata processes data at the edge, keeping it close to the source. The lightweight open-source Netdata Agent acts as a distributed database, enabling the construction of complex observability pipelines with modular, Lego-like simplicity.
Netdata provides A.I. insights for all monitored data, training machine learning models directly at the edge.[8] This allows for fully automated and unsupervised anomaly detection, and with its intuitive APIs and UIs, users can quickly perform root cause analysis and troubleshoot issues, identifying correlations and gaining deeper insights into their infrastructure.
Starting with v1.12, Netdata collects anonymous usage information by default and sends it to Google Analytics, a feature which can be disabled via manual configuration.[9]
A 2023 peer-reviewed empirical study evaluating monitoring tools on Docker-based microservice systems found that Netdata was the most energy-efficient tool among those evaluated, while having no statistically significant impact on execution time under the tested workloads.[10]
Operation
When executing the daemon on Linux using the netdata command, threads are generated that collect information from each resource, using internal and/or external plugins. In turn, it keeps a record of the values collected. Depending on configuration, Netdata stores collected metrics either in memory or on disk using its database engine. [11]
It operates as a stand-alone web server for its own static files, necessary for the representation of its dashboards.[12][13] It provides a REST API so that the browser can access the information.
Each installation of the application works autonomously. Although different running instances of the application can be saved to one dashboard, every Netdata instance is independent. Only the browser can connect all installations of different systems, unifying graphics from different sources as if they came from the same server.
Development
Netdata is currently maintained by nearly 400 contributors,[14] all helping (at various levels) to serve the thousands of individual users and businesses[15] who utilize this tool.
The user with the most contributions is currently Costa Tsaousis, the CEO and Founder of Netdata, with over 600,000 additions to the code.[16][17] The second most-active user is Ilya Mashchenko.[18]
The all-time most popular addition to Netdata appears to be adding support for data collection from Vnstat, a pull request by Noah Troy with nearly 200 individual comments (more than any other pull request).[19]
The all-time most popular feature request appears to be adding support for running multiple freeipmi jobs from the same Netdata.[20]
This page is based on this Wikipedia article Text is available under the CC BY-SA 4.0 license; additional terms may apply. Images, videos and audio are available under their respective licenses.
