Skip to content

Zabbix items and monitoring data collection

In [[Zabbix]], monitoring data is fundamentally organized around items. An item represents an individual metric or piece of data collected from a host, such as CPU usage, disk space, or network traffic statistics.^[600-developer-operation-maintenance-zabbix-zabbix-intro.md]

Item Configuration and Storage

Detailed information about every item is stored in the items table^[600-developer-operation-maintenance-zabbix-zabbix-intro.md]. Each item is defined by several key parameters:

  • Unique Identifier: The itemid serves as the unique identifier for each collection item^[600-developer-operation-maintenance-zabbix-zabbix-intro.md].
  • Host Association: Items are linked to specific hosts using the hostid^[600-developer-operation-maintenance-zabbix-zabbix-intro.md].
  • Collection Interval: The delay parameter specifies the frequency at which data is collected^[600-developer-operation-maintenance-zabbix-zabbix-intro.md].
  • Retention: The history field defines how long the item's collected data should be stored^[600-developer-operation-maintenance-zabbix-zabbix-intro.md].
  • Status: The status field indicates the current state of the item (e.g., a value of 0 typically denotes an active, enabled item)^[600-developer-operation-maintenance-zabbix-zabbix-intro.md].
  • Units and Naming: Items are configured with a specific name and units of measurement^[600-developer-operation-maintenance-zabbix-zabbix-intro.md].

Data Types and Storage

Zabbix uses a dedicated set of tables to store the actual data collected by items, categorized by data type^[600-developer-operation-maintenance-zabbix-zabbix-intro.md]. The history tables (such as history, history_str, history_log, and history_uint_sync) store raw data, with different tables handling different data types (numeric, string, log, etc.)^[600-developer-operation-maintenance-zabbix-zabbix-intro.md].

These raw history tables serve as the data source for short-term graphing, such as plots covering one hour or one day^[600-developer-operation-maintenance-zabbix-zabbix-intro.md].

For long-term storage and analysis, Zabbix utilizes trends tables (like trends and trends_uint)^[600-developer-operation-maintenance-zabbix-zabbix-intro.md]. Unlike the raw history tables, these tables store processed statistical data^[600-developer-operation-maintenance-zabbix-zabbix-intro.md].

The system calculates trend statistics periodically (typically every hour) based on the raw history data^[600-developer-operation-maintenance-zabbix-zabbix-intro.md]. These statistics include the minimum, maximum, and average values over the statistical interval^[600-developer-operation-maintenance-zabbix-zabbix-intro.md]. This allows for efficient data storage and visualization over longer time ranges.

Items are often grouped together for visualization purposes. The graphs table contains user-defined chart configurations, while the graphs_items table maintains the relationship between a graph and the specific items included in it^[600-developer-operation-maintenance-zabbix-zabbix-intro.md].

Sources

^[600-developer-operation-maintenance-zabbix-zabbix-intro.md]

  • [[Zabbix]]
  • [[Time-series data]]
  • [[Database schema]]