At approximately 4:00 PM CT on 11/09/2015, we updated the filtering functionality in LeadConduit Classic, inadvertently causing a change in the way it works, and causing some lead delivery problems. Full resolution was made today at approximately 9:00 AM CT.
A change in one of the libraries we use caused the filter behavior to change. When a filter blocked a lead, it resulted in an 'error' outcome rather than a 'failure' outcome.
The issue was not uncovered until approximately 8:30 AM CT on 11/10/2015, at which time we were able to identify the issue and patch the code.
If you use filtered deliveries, the impact to you depends on the configuration of the affected destination's "on delivery error" setting under the multi-destination behavior.
If it's set to “do not continue to the next destination, retry later”, then the process is halted and retried a few minutes later. If the error occurs again, then the process repeats until the error resolves.
If it's set to the default, "continue to next destination, retry later", then that destination will be skipped and the remainder of the destinations will be processed. Later, that destination will be retried. If the error occurs again, then the destination is retried until the error resolves.
If your deliveries have alerts that disable destinations after a certain number of consecutive errors, then this issue could have also triggered that.
We apologize for the inconvenience, and are continuing to look into what happened and the oversight related to this release in an effort to prevent similar situations in the future.