Our planners are running into errors when uploading forecast CSVs that include the Daylight Saving Time “spring forward” transition (e.g., March 8, 2026). Specifically, any rows that include times in the nonexistent hour (e.g., 2:00–3:00 AM local time) cause the upload to fail with confusing or generic validation errors.

Right now, the behavior isn’t obvious to users, and it’s not clear from the UI that DST is the root cause of the issue.

The CSV parser should detect DST transition gaps and either skip nonexistent times, provide a clear DST-specific error message, or parse uploads in UTC to avoid these failures.