For a true distributed environment, we should allow zabbix_server to communicate with multiple zabbix_proxies and perform item checks from multiple checkers.
Ideally simply because icmpping says the host is down, does that really mean its down? or is there just some temporary routing problem.
It would be nice to detect that:
2 / 4 locations cant reach host 'A'
Location 1 has 20% packet loss to host A
Location 2 has 50% packet loss to host A
Location 3 cant reach host A
Location 4 cant reach host A
Once this is available -- zabbix will truely be the "distributed monitoring" king. Currently im just taking multiple zabbix servers and aggregating the results together... But this its not so fun.
Ideally simply because icmpping says the host is down, does that really mean its down? or is there just some temporary routing problem.
It would be nice to detect that:
2 / 4 locations cant reach host 'A'
Location 1 has 20% packet loss to host A
Location 2 has 50% packet loss to host A
Location 3 cant reach host A
Location 4 cant reach host A
Once this is available -- zabbix will truely be the "distributed monitoring" king. Currently im just taking multiple zabbix servers and aggregating the results together... But this its not so fun.