Changing Dataiku's timezone for DSS formula's

[SanderVW]

How do you adjust the timezone that is used for the DSS formula language? For instance, the now() formula is now 2 hours behind (was 1 hour prior to summertime). It's not necessary to make any retroactive changes, we would just like to get our local time from now on. 

I imagined this would be a global setting somewhere in the instance settings but have yet to find it.

[tgb417]

@SanderVW ,

From my point of view this is a long running conversation with the Dataiku team. In general they enforce a practice of storing and managing data times in Dataiku datasets in GMT. (Which I don’t think goes through daylight saving time craziness. So that’s a bit helpful).

That said they do not provide very robust mechanisms to show date times in local time zones. For doing things like graphs or easily understanding your datasets by looking at them can be challenging.  In my case I’m in the eastern time zone in the United States. And my times are always either 4 or 5 hours off as I look at them on the screen, and knowing how to think about data in the spring and fall can be particularly challenging. 

Some times I will try to create a column in my local time. But that has generally been an ugly solution. 

[Turribeach]

Dataiku takes the system time when it starts. For instance to set my DSS instance to UK time zone I added this to ./bin/env-site.sh:

export TZ=GB

[tgb417]

@Turribeach 

Is that documented anywhere?   I’m wondering about side effects of changing this value, with existing projects and on existing servers and with existing datasets.

Does this change the way data is stored. Just how it is displayed?  How various visual recipes work?  How Python Recipes work? Existing log files.  Statuses on datasets from past updates….

Right now my instance is in default configuration. If I change this what other things will happen?