Ignore timezone shifts for histogram labels

When a user requests a visualisation grouped by the start/end date of a process instance, the labelling for each bucket can be incorrect if the server timezone is before (or East of) that of the client. This can be observed for any interval, so the data that should be in the February bucket could instead be labelled as January, for example. We need to research the best solution to this as currently the timezone of both the client and server is not configurable in Optimize and there are pros and cons of different solutions.

• If we used the server timezone everywhere, the client could potentially see data labelled as in the future of their timezone, which would be confusing. There would need to be some changes to how we make this visible to the user. If we went down this route of using server timezone, one option would be to at least make this configurable rather than using system default.

• If we adjusted data based on the client timezone or allowed users to have their own timezone configured, then different users within an organisation would see different data for the same reports depending on what their timezone is set to, which could be even more confusing for sharing etc. We also have no current way to store these user timezone preferences given Optimize not doing user management.

To demonstrate the potential problem, you can do the following to reproduce:

Given: A process instance report with a process instance view, grouped by either start or end date of process instance by any interval.

When: The client timezone is before the timezone of the Optimize server.

Then: The label is incorrectly shifted according to the number of zones different.