I am trying to log Lambda function failures in such a way that event and context information is saved so that the event information can, if deemed necessary, later be manually republished to the function's trigger. I do not want to handle this logic in the functions themselves.
What I've tried so far:
- CloudWatch alarms on the error metric. They tell me only that a function has failed.
- Looking in the CloudWatch logs. I only see the coded failure messages emitted from each function.