There are different exit codes 0-5 which cause different behavior (not documented in the manual, but in the example scripts), to display error or warning messages and also abort:
There is no need to change the database, it already contains the result logs (what you see when you go to show log on the backup). It would only also have entries when the job failed due to exceptions, which is more intuitive in my opinion.
I also looked in the code and results are only ever written to the database log, never parsed. This means we can add any fields or enum values we like and backwards compatibility is unaffected.
It could still be used if you have an existing script that returns a different code, e.g. exit 1 to fail which is quite common.