SAT Testing confirms that a crash produces at least one AirbyteTraceMessage

[pedroslopez]

Full context in parent issue #12322

The SAT should verify that an AirbyteTraceMessage is emitted when something goes wrong.

This may be done by using the invalid_config.json that most connectors have and running a sync with it.

[Phlair]

Open Q:

• Should we actually be testing for this is SAT if the implementation described here/here means that connectors will all be doing this automatically? In other words, this seems like a concern of the cdk implementation rather than the individual connector...
  • If we do want to have this test in SAT, then we should also create a similar issue for DAT.

[Phlair]

notes from backlog grooming:

• add this as a default test, opt-out possible for cases where cdk not being used.
  • document how to opt-out
• clarify stance on AirbyteTraceMessage in protocol
  • strict requirement? Highly recommend? Optional?
• create issue for DAT
• add dependencies on cdk/java changes