The logs are great for identifying errors and the volume. But without seeing the full payload, seeing that a response code of 400 occurred is telling us only half of the problem.
I’m happy to be directed to somewhere that this is viewable.
I also pre-empt the “You should handle the response code and reporting better on your side” response. However, many of us may be the Ops people and not the development / engineering team, and often need proof that there is a problem to be investigated