We are on version 1.8.3 and having lots of false alerts when checking time sync (ntp) on hosts that are on remote proxies.
I think the problem is the Zabbix Server processes the time via fuzztime() when it receives the data from the proxies. But our proxies are all over the world, so if our Hong Kong proxy's data is delayed 15 minutes due to Internet connectivity, then we get alerts on all the servers since the time is now different.
Any ideas ? Ideally we check the time vs. the lastclock/gather time, not Zabbix server receive time.
Item: system.localtime
Trigger: {Template_Linux:system.localtime.fuzzytime(180)}=0
So if the data was gathered at 15:00:00 and local time on host was 15:00:01 but the data arrives at Zabbix at 15:05:30, we'll get a trigger on this since 15:00:01 <> 15:05:30
I think the problem is the Zabbix Server processes the time via fuzztime() when it receives the data from the proxies. But our proxies are all over the world, so if our Hong Kong proxy's data is delayed 15 minutes due to Internet connectivity, then we get alerts on all the servers since the time is now different.
Any ideas ? Ideally we check the time vs. the lastclock/gather time, not Zabbix server receive time.
Item: system.localtime
Trigger: {Template_Linux:system.localtime.fuzzytime(180)}=0
So if the data was gathered at 15:00:00 and local time on host was 15:00:01 but the data arrives at Zabbix at 15:05:30, we'll get a trigger on this since 15:00:01 <> 15:05:30