Hello,
We've got a batch of about ~300 monitored computers that reset themselves between 3:05 to 3:35 AM semi-randomly, within 5 minutes intervals (some at 3:05, some others at 3:10... to 3:35). The uptime for those computers is monitored by a cloned Windows by Zabbix Agent template, with the default custom interval of 1 minute for 1-7,00:00-24:00. To avoid being warned by these resets we added an exception (interval of 0 for 1-7,02:55-03:55), because they meant more than 600 e-mails for each one of us. And this was working fine until we updated one of the last minors on Zabbix Server 7 LTS (can't remember if that was 7.0.8 or 7.0.9).
Without having changed the custom interval during months, we recently noticed that we're being e-mailed with the reset warning again (uptime < 10m). Trying to debug the problem, we noticed that we're receiving data during the time where the interval 0 shall be applied and the interval 0 is applied between 03:55 and 04:55, shifted one hour forward.
We reviewed local time and NTP configuration in both ends (Zabbix Server and the monitored computer), not seeing any mismatch nor misconfiguration. Actually, they are synchronized with the same NTP server and the data timestamp is correct. We also checked that there is no summer time application, both of them are on CET. But it seems more like an issue with the template itself.
Did anyone encounter with this situation? My solution (temporary) was to shift back one hour the definition of the interval.
Thanks. Best regards,
Oriol
We've got a batch of about ~300 monitored computers that reset themselves between 3:05 to 3:35 AM semi-randomly, within 5 minutes intervals (some at 3:05, some others at 3:10... to 3:35). The uptime for those computers is monitored by a cloned Windows by Zabbix Agent template, with the default custom interval of 1 minute for 1-7,00:00-24:00. To avoid being warned by these resets we added an exception (interval of 0 for 1-7,02:55-03:55), because they meant more than 600 e-mails for each one of us. And this was working fine until we updated one of the last minors on Zabbix Server 7 LTS (can't remember if that was 7.0.8 or 7.0.9).
Without having changed the custom interval during months, we recently noticed that we're being e-mailed with the reset warning again (uptime < 10m). Trying to debug the problem, we noticed that we're receiving data during the time where the interval 0 shall be applied and the interval 0 is applied between 03:55 and 04:55, shifted one hour forward.
We reviewed local time and NTP configuration in both ends (Zabbix Server and the monitored computer), not seeing any mismatch nor misconfiguration. Actually, they are synchronized with the same NTP server and the data timestamp is correct. We also checked that there is no summer time application, both of them are on CET. But it seems more like an issue with the template itself.
Did anyone encounter with this situation? My solution (temporary) was to shift back one hour the definition of the interval.
Thanks. Best regards,
Oriol
Comment