https://www.zabbix.com/integrations/rabbitmq
All four templates for RabbitMQ contain a trigger that is based on a master item.
Data collection is done using dependent items.

This is against our best practices which suggest not keeping historical data for master items. Instead, the triggers should be put on dependent items to save DB space and not store any history for the master item.

In the case of production environment monitoring for RabbitMQ, the data collected over an hour, which is currently used as the default history period, can be quite large.

Something similar can be seen also on Apache by HTTP/ by zabbix agent, Nginx by HTTP/by zabbix agent, and probably also a few more templates.