Possible wrong event's time value in "Events"

[yugo45]

Good morning,

I have some strange problem with event's time in "Events": time of the event "ОК" earlier (!!!) than "PROBLEM".
Event turns into status "PROBLEM" at 06.05.2018 03:18:00 , and turns into "ОК" - at 06.05.2018 03:15:46

It's critically for SLA report, because SLA still counting that the service is still in "PROBLEM".
,
Any advice is greatly appreciated.

Zabbix 3.0.1

[aigars.kadikis]

Hello, yugo45,

In what way these metrics has been gathered? Are we talking about Item type: "Zabbix agent (active)"?
Did any other servers in this time period has OK state later than Problem state?

Regards,

[yugo45]

Hello!
Metrics has been gathered by 3rd-party solution, not by Zabbix agent.
No, problems occur on different servers at different times. Not at the same time.
(unfortunately, I do not know how to upload screenshots in good quality in forum. I used other service)

http://i104.fastpic.ru/big/2018/0514/c0/3c96835c225c164212cb48dfc6acb0c0.png

http://i104.fastpic.ru/big/2018/0514/b7/_d4472272745337c2040882b3d9500db7.png

http://i104.fastpic.ru/big/2018/0514/f5/8c6c72536333df93d4d58b89b10dbcf5.png

http://i105.fastpic.ru/big/2018/0514/f9/4993ec1e400a07d30e38c67783a5a5f9.png

[aigars.kadikis]

Hello!
Metrics has been gathered by 3rd-party solution, not by Zabbix agent.
No, problems occur on different servers at different times. Not at the same time.
(unfortunately, I do not know how to upload screenshots in good quality in forum. I used other service)

http://i104.fastpic.ru/big/2018/0514...dfc6acb0c0.png
http://i104.fastpic.ru/big/2018/0514...b3d9500db7.png
http://i104.fastpic.ru/big/2018/0514...89b10dbcf5.png
http://i105.fastpic.ru/big/2018/0514...7783a5a5f9.png

Please expand your story. How do you deliver these metrics to Zabbix DB? Is it by using "Zabbix trapper" or "SNMP Trap" or another channel?

Regards,

[aigars.kadikis]

It looks like a network outage (from 14:33:30 to 14:38:30) here because your trigger fires up if there is no data in last 5 minutes, but 'Latest Data' represents that there is data in this time interval.
Whatever service pushes this data to zabbix server, it has a cache big enough to store some metrics offline and deliver those to the server when possible.

As an workaround you can switch this trigger severity to level 'Information' and it should be not included in SLA.

Regards,

[yugo45]

Aigars, thank you for your help in solving the problem! It's network and performance issue for server that pushes data to zabbix. Solved.