Optimizing Zabbix Monitoring for Linux and Windows Systems

Zabbix uses its built‑in agent to monitor OS metrics; after installing the agent, templates can be linked in the frontend, and custom items can be created with system.run[command,<mode>] when needed.

The article explains the selection and optimization of OS‑related items for both Linux and Windows, covering agent‑type keys, passive vs active modes, and when to use each based on environment size.

For Linux, it recommends setting most discovery rules to passive mode, adjusting polling intervals (e.g., 1 min for CPU/memory, 1 h for informational items), disabling unused triggers, and using macro variables such as {$CPUIOWAIT} to fine‑tune thresholds.

It also shows how to use LLD macro variables, e.g.,

{host:vfs.fs.size[{#FSNAME},pfree].last()}<{$LOW_SPACE_LIMIT:"{#FSNAME}"}

, and how to set host‑level macros to override template defaults.

Memory monitoring keys are listed with explanations (total, free, active, etc.), and the article advises using pavailable for alerts.

File‑system items are streamlined to three key metrics (total, used, percent used) to reduce load.

For Windows, the guide suggests disabling the default service discovery, enabling it per host when needed, and using the perf_counter[<counter>,<interval>] and

eventlog[<name>,<regexp>,<severity>,<source>,<eventid>,<maxlines>,<mode>]

keys to monitor performance counters and event logs, with a brief description of their parameters.