Description
Currently, we ad-hoc handle timezones as needed, but there's not a clear standard for NPS-ActiveSpace as a whole.
One straightforward approach is to convert all datetimes to timezone-aware UTC timestamps on ingest.
Use Case
Using UTC for all timestamps would allow internal library functionality to confidently assume timestamp format across the codebase. Since we potentially deal with data from monitoring sites across different timezones, we should have clear logic and standards around how we handle this data to avoid timezone-related bugs.
For end-user display, we could convert back to local timestamps at the boundary.
Proposed Solution
On ingest, convert all timestamps to UTC. Python datetime types and Pandas timestamp columns should be made timezone aware. There should be straightforward functionality to convert these temporal types to UTC since storing timestamps as UTC is a standard approach.
Alternatives Considered
No response
Description
Currently, we ad-hoc handle timezones as needed, but there's not a clear standard for NPS-ActiveSpace as a whole.
One straightforward approach is to convert all datetimes to timezone-aware UTC timestamps on ingest.
Use Case
Using UTC for all timestamps would allow internal library functionality to confidently assume timestamp format across the codebase. Since we potentially deal with data from monitoring sites across different timezones, we should have clear logic and standards around how we handle this data to avoid timezone-related bugs.
For end-user display, we could convert back to local timestamps at the boundary.
Proposed Solution
On ingest, convert all timestamps to UTC. Python
datetimetypes and Pandas timestamp columns should be made timezone aware. There should be straightforward functionality to convert these temporal types to UTC since storing timestamps as UTC is a standard approach.Alternatives Considered
No response