Netdata vs. Zabbix: The Open-Source War Over Tracing Server Load Down to the Executable
The core problem is establishing a reliable, historical record of which processes are actually driving system load over time on Linux servers.
The technical argument is deeply split. Some users, like mosiacmango, champion Netdata as a simple, FOSS solution for local metric logging. Others favor established, battle-tested SNMP frameworks like Cacti, Icinga, or Zabbix, which provide broad monitoring scope. A major point of contention is modern 'AI' monitoring; 'vk6flab' expressed profound mistrust, warning that proprietary ML tools are suspect due to potential hidden functionality and an inability to gather continuous training data.
The community gravitates toward robust, established tools. While Netdata gained traction, the general leaning favors systems (like Zabbix or SNMP stacks) capable of comprehensive, configurable data collection, while simultaneously acknowledging the fundamental difficulty in tracking specific executable load changes due to rapid OS and application updates.
Key Points
Netdata is recommended as a simple, FOSS alternative for local system metrics.
mosiacmango gave Netdata high praise for its usability.
SNMP-based systems (Cacti, Icinga) are seen as robust for monitoring loads, memory, and NIC usage.
rowinxavier confirmed load attribution is 'definitely manageable' with these stacks.
Skepticism targets proprietary 'AI' monitoring solutions.
vk6flab voiced strong reservations about ML tools, citing concerns over opaque functionality.
atop and sar represent specialized or historical tooling options.
raoul pushed atop as a daemon option, while daqu recalled the reliability of 'sar'.
Zabbix is noted for its flexibility across disparate monitoring sources.
Oisteink praised Zabbix for integrating SNMP, IPMI, REST APIs, and custom agents.
Source Discussions (3)
This report was synthesized from the following Lemmy discussions, ranked by community score.