"Successful" puppet agent reports hide provider errors

It's possible to get a successful agent run report that on closer inspection include errors.

For instance, an error points to a failure with the Chocolatey provider. Because the provider failed out, it did not return a complete list of resources to manage, and the agent run completes successfully on this smaller catalog without raising an alarm.

There is no way for an administrator to notice this problem and correct it without drilling down into green agent reports to confirm whether an error has occurred or not. On large scale deployments this solution rapidly becomes unmanageable.

Agent runs need to return some condition when providers fail that highlights that this has occurred. There needs to be a scaleable way to monitor and report on this kind of failure.

Please add all errors and warnings for consideration as failed run status assignment criteria here.