Timezone issues with trigger actions

[slyper]

Hello Zabbix Forum,

We run a Zabbix server (zabbix-server-pgsql:latest) and frontend (zabbix-web-apache-pgsql:latest) container in Amazon ECS.
We configured some trigger actions containing a time period.
After working hours, the monitored servers are shutdown, and we don't want to get notified for problems during night time (e.g. ICMP ping fails).
The time period we use for this in the trigger action: 1-5,09:00-19:00

The problem is, we still get these ICMP ping failure notifications:


In the notification, you can see the local time is 19:03 (3 minutes after the server is shutdown), but Zabbix thinks it's 18:03 (hence the notification).

Some more details:
- the frontend container is run with the environment variable PHP_TZ = Europe/Brussels
- the GUI default timezone setting is also Europe/Brussels
- the local and server time widget seem to indicate the same time

- while on a bash terminal in the zabbix server container, UTC is shown (so same date, only different timezone)


How can I solve this issue? Do I need to use UTC times for my time period? This would mean I would need to adjust them when switching timezone (daylight saving time).

kind regards

[slyper]

Bump
We solved this by using UTC times in the trigger action's time period. This is not our preferred solution, as our servers change timezone 2 times a year (CET and CEST), so we'll need to update the times in Zabbix as well.
Anyone knows a better solution?

[Markku]

Is there a reason you cannot fix the timezone in the Zabbix server container? I mean, if Zabbix is running on UTC (that doesn't use DST) and your alerting times depend on local time (that uses DST), you will have one-hour-off problems at some point.

Markku

[KopyrightKid]

I also have run into this at my organization - if anyone has found a fix other than using UTC in trigger conditions and just keeping an eye out for daylight savings time I'm all ears!

[KopyrightKid]

Update: We opted to set the host machine to our local timezone