We currently use PRTG, but it has become difficult to maintain and the price increase means we will not renew. We need a free or open-source replacement for fairly basic monitoring rather than an especially complex setup.
The main requirements are Jira ticket creation, SSL certificate expiration alerts, notifications when network devices or servers go offline, sustained high CPU alerts on servers, and possibly warnings when a switch port drops from 1 Gb to 100 Mb.
Our environment includes Cisco Catalyst 3000-series and Nexus switches, a voice router, FortiGates at the main and branch offices, FortiSwitches, FortiAPs, FortiAuthenticator, FortiAnalyzer, a Nimble SAN, HPE servers running VMware, mostly Windows servers, and a few CentOS systems. I am new to monitoring platforms and am considering Zabbix, Prometheus, LibreNMS, or similar tools. Which option would be the most practical for this use case?
4 Answers
For Zabbix, start small instead of trying to monitor every metric immediately. Deploy it with the official container setup or packages, add a few representative switches, firewalls, and servers, then tune availability, CPU, certificate, and interface triggers before expanding. Jira integration can be done through a webhook, so it should not determine the platform choice. The open-source edition is usable without feature paywalls, but the real cost is the time needed to learn and maintain the system.
Checkmk can also cover this type of mixed environment and has strong host and service discovery. It is another good choice if you want both infrastructure and server checks, although hardware support and the exact features available should be verified for your devices. I would not choose Prometheus as the primary tool here: it is excellent for application and time-series metrics, but network gear often requires more manual SNMP exporter configuration.
The main decision is probably whether you want the fastest network-only deployment or one platform that can grow into detailed server monitoring. LibreNMS favors the former, while Zabbix or Checkmk favors the latter.
Zabbix is probably the best all-around fit. It has a learning curve at first, but the day-to-day administration becomes straightforward once the concepts click. It supports agents and SNMP, and there are ready-made templates for common Cisco and Fortinet devices. Your listed requirements can be handled with templates, triggers, and webhooks, including Jira integration. Plan on spending several days to a week learning and tuning it rather than expecting a completely instant setup.
That makes sense. The initial learning curve is what made me hesitate, but having templates for the network equipment would help a lot.
LibreNMS is worth considering if the priority is getting network monitoring running quickly. It auto-discovers many SNMP devices and can cover switches, firewalls, server management interfaces, interface changes, availability, and CPU alerts with relatively little configuration. It is generally stronger for network equipment than for detailed server monitoring. SSL certificate monitoring may need to be supplemented with another small tool, and configuration backups can be paired with something like Oxidized.
It is especially convenient when most of the environment exposes useful SNMP data. For deeper Windows and Linux monitoring, Zabbix is usually the more complete single platform.
That phased approach sounds manageable. I will probably build a small test deployment first and validate the alerts before migrating everything.