Add trace logging and request data in logs when errors occur

[erezmus]

Follow up from (#2009)

We currently are having random unexplained failures in the pipelines and the logs don't seem to give enough information as to the reason.

Proposal is to add:

• Stack trace
• Dump of request data