The issue originally described in bug ZBX-19481 is still present in Zabbix 5.0.14

It seems that the issue is not solved.

I updated my Zabbix server to version 5.0.14 but still get messages "... has just been restarted" although this planned restart has been executed inside the maintenance period.

Please double check that Zabbix server was restarted after upgrade, is it possible that alerts were generated on 5.0.13 or are those alerts from new maintenance on 5.0.14 ?

Same for version 5.4.3. Updated on 27th of July to this version but still got notifications within maintenance windows.

@Vladislavs Sokurenko: By server restart you mean the whole system, or just the zabbix services?

Originally was experiencing the same result as tni, however, after following Vladislavs' suggestion of a server reboot I can confirm that we're no longer experiencing the issue as we did on ZBX-19481

I would also add that we updated and rebooted all proxy's too, not sure if that was required or not but would assume so given that much of the trigger logic etc has moved to the proxies.

Please double check that Zabbix server was restarted after upgrade

Yes, the server was restarted after the update. To be absolutely sure I rebooted the server once again, but the issue persists.

??is it possible that alerts were generated on 5.0.13 or are those alerts from new maintenance on 5.0.14 ???

these are new alerts, they are generated daily.

I did investigate this issue further and found out that something strange with timestamps/timezone is going on.

The item data which is causing an alert to be triggered is shown in graphs at the correct time. The alert email is sent at the correct time when the trigger event happens. But the time shown in the email body i.e. "Problem has been resolved at 00:23:23" is off by -2h (which correlates to our timezone offset of +2h).

I further recognized that there also is a time offset when the maintenance period actually begins. I.e. a maintenance period that is configured to start at  12:00 starts in reality at 14:00

The timezone is and always was correctly set in /etc/php-fpm.d/zabbix.conf  and did work fine when I was still using version 5.0.2, but not since the upgrade to 5.0.13 and then 5.0.14

I found out that the servers timezone was set to UTC while the timezone in  /etc/php-fpm.d/zabbix.conf  was CEST.

Setting the servers timezone to CEST solved the issue. The time shown in the email body and start time of the maintenance periods is correct now.

With version 5.0.2, having the servers timezone set to UTC worked. This leaves the question what changed between version 5.0.2 and 5.0.13 / 5.0.14 that changed this?

I'll add my observations from today on version 5.4.3 (timezone of the server is EEST, but some of its users are in timezone CEST, therefore configuration below). Having this maintenance configuration:

and action configuration:

events were tagged:

but action still sending e-mails ... maybe mistake on my side, maybe some quoting needed somewhere , but it is also possible, that this is related to this ZBX-19720 somehow ...

(during testing, I had At 13:00 in maintenance period schedule to hit maintenance window with the test)

Closing as cannot reproduce, please feel free to reopen if issue still persists.