Use this guide as an implementation reference. Validate commands, policies and version-specific behaviour in a lab before applying changes to production systems.
What to monitor
Start with availability, CPU load, memory and swap, filesystem capacity, disk I/O, network interfaces, uptime and critical services. Add application-specific metrics only after the basic health signals are reliable.
Architecture overview
A typical deployment includes a Zabbix server, database, web frontend and monitored hosts. The Zabbix agent collects host metrics. Agent 2 supports additional plugins. In active checks, agents connect to the server or proxy; in passive checks, the server or proxy polls the agent. Proxies are useful for remote sites and segmented networks.
Install and register an agent
Install the agent package using the official repository and instructions for your Linux distribution and Zabbix version. Configure the server address, hostname and connection mode in the agent configuration file. Example settings:
Server=10.10.50.10\nServerActive=10.10.50.10\nHostname=web-server-01Use your actual Zabbix server or proxy address. Restrict agent access at the host firewall to trusted monitoring systems.
Add the host in the web interface
- Create or select a host group.
- Add the host using the exact configured hostname.
- Set the agent interface address if using passive checks.
- Link an appropriate Linux by Zabbix agent template.
- Confirm the interface is available and latest data is arriving.
Triggers, thresholds and notifications
Templates provide discovery rules, items and triggers. Review default thresholds against the server role and normal baseline. A 90% disk threshold may be reasonable for one workload and too late for another. Configure action conditions, notification media and escalation paths, then test delivery.
Dashboards and operational practice
Build dashboards for availability, capacity, problem status and service health. Use tags and host groups to filter by environment, location and service owner. Keep alert messages actionable: identify the host, symptom, severity and runbook link.
Security and troubleshooting
- Use supported TLS encryption for agent communication.
- Limit allowed agent sources and avoid exposing monitoring ports publicly.
- Keep Zabbix server, agents and database patched.
- Check DNS, routing, firewall rules, time synchronization and hostname consistency when data is missing.
- Review server and agent logs and verify template compatibility.
Conclusion
A useful monitoring system turns raw metrics into timely operational decisions. Begin with essential health and capacity metrics, tune alert thresholds to real workloads, secure the monitoring path and continuously improve dashboards and runbooks.
