Via @mikand13 on Gitter, the documentation about configuring custom health HTTP status mappings may give the impression that the custom mapping is in addition to the default mappings. It is actually in place of the default mappings. This is intentional as it allows users to replace the defaults but the documentation is unclear.
We should review and update this section to clarify things.
Comment From: mikand13
Correct, somewhere around here would be nice:
"You might also want to register custom status mappings if you access the health endpoint over HTTP. For example, the following property maps FATAL to 503 (service unavailable) (...)"
Thanks for looking into it
Comment From: wilkinsona
Closing in favour of #20716.