Having explored basic monitoring with a Zabbix agent before, we looked at a major agentless monitoring solution in this chapter—SNMP. Given the wide array of devices supporting SNMP, this knowledge should help us with retrieving information from devices such as printers, switches, UPSes, and others, while also listening and managing incoming SNMP traps from those.
Beware of starting to monitor a large number of network devices, especially if they have many interfaces. For example, adding 10 switches with 48 ports, even if you monitor a single item per switch once a minute only, will make Zabbix poll eight new values per second (480 ports once a minute results in 480/60=8 new values per second). Usually, more values per port are monitored, so such an increase can bring a Zabbix server down and severely impact network performance even when SNMP bulk get is used.
While we have created several hosts by now, we only paid attention to the host properties that were immediately useful. In...