RFC: fix: FatalError set reference to causing error

[ivanjh]

• All tests pass (no - I don't have suitable tokens to test all integration tests)
• I have run npm run doc (not part of external api)

Description

FatalError does not provide enough context for understanding underlying issues.
Where available, attach the "causing" error to the FatalError to allow later inspection on .cause (see https://nodejs.org/api/errors.html#errorcause)

Serializing Error objects with util.inspect() recursively serializes error.cause if it is set and would allow better log messages in consuming applications.

I don't believe the existing approach of copying of the stack from the causing error is "least surprise" for consuming authors, but I do not yet understand reasons for why it is as is, or likelihood of breaking existing code.

References

Recent node versions introduced a bug (nodejs/node#54359) in certain network configurations that is becoming more numerous triggering EFATAL errors.

Reports often miss much depth, with only "EFATAL: AggregateError" as a description:

• Issues at this repo: Stop listening for new message events and crash app #1174
• Issue on consuming repo: getting a polling error: aggregate error  windkh/node-red-contrib-telegrambot#345

[windkh]

+1

[tedtribe]

+1