OMD does look good - and I do have the know-how and experience to handle Nagios / iCinga.
However, one of the main arguments for Zabbix and against most Nagios / iCinga and others:
It comes from one company, is in one release tarball. No dependencys on other projects or tools.
Ever try to get a grafik showing load of a server or network link from Nagios? Yes, you NEED 3rd Party Components. Grafana, NagVis, et al. Loads of third party dependencys. When sh*t happens, everyone passes the blame to someone else, no one actually fixes it.
I tried for hours to get Zabbix to make a graph, only then did I realise the grapf was there all along, out of the box as default. The only additions are Test-Templates, which can be third party. OK, the prefered way is using agents (available for Windows, Linux/UNIX and Mac), but SNMP monitoring works extremly well.
The testing was clinched, when a test environment of Zabbix showed me an available Update on a 2010 Synology (Running only to make 2nd generation in house saves of the new Synology).
I was never able to get that much information out of Nagios / iCinga...
Mind you, I even had my Nagios reporting Phone Calls, VoIP Info and the usual System data via SNMP from my Snom SIPfones...
All Information and events are stored in a MySQL or MariaDB Database, so for backup you only need a Tarball of the current directory and the Database dump.
I'm running Zabbix now in a vm on proxmox, but I prefer to have it integrated in Nethserver or whatever my clients are using (I have some running a Mac Server). Some clients do not have a vm environment but use only one Server (Linux or Mac) and a NAS (mostly Synology, some older Thecus) for Infrastructure. Zabbix would work in these environments well too.
But OMD is still an interesting collection / tool to have in one's library...
My 2 cents...
Andy Wismer, from Switzerland