Last modified: 2013-06-18 16:59:08 UTC
If a queue (ActiveMQ) outage happens then the interface will send back a cryptic message informing the donor that we've failed pushing the message to our internal queue. This is both hard to understand and mostly useless to the donor. Instead it would be nice to fail gracefully by possibly writing the transaction locally and replaying it to the queue later. We should have some kind of monitoring in place to track local queue growth in case of a large unplanned outage. At no part of the pipeline should we message our donors a message that makes no sense to them. If a donation has been accepted through payflow then we should always show the donor our thank you page.
This has been fixed.