These reports also should be able to capture relevant error snapshots and perhaps even other event information (like traces) so that we can provide contextual information for the time period to go along with the metric data that is already presented.
Customers have asked for such a feature, though the paradigm for integration is a little shaky. We'd need to consider a method by which we could choose which traces/errors were integrated. Blindly including all from an agent may create a 500-page report. However, I agree that it would be really nice to have the ability to share, for example, all the SQL errors (dupicate keys, constraint violations, etc.) that are occuring into a PDF that I could send back to a dev team. It would be highly helpful if we can get the format correct.