The best solution is to compute the SLA timeline prior to loading data into GD platform and display it as date attribute rather than metric.
Workaround can be to use attributes like “Day of Month”, “Month” and “Year” and use some complex formula to add 4 days and get the result in form of number with format MMDDYYYY (e,g, 03082022 for March 8, 2022) and then use metric formatting to display it as 03/08/2022. Note that formula needs to be complex because each month has different number of days. Maybe you may consider to use some formulas from
https://en.wikipedia.org/wiki/Julian_day or
https://en.wikipedia.org/wiki/Ordinal_date to extract day and month components from ordinal number of date.