Lots of ways. You can wrap a particular operation (say, calling a .NET component) in a \”scope\” shape marked as a long running transaction. Then you can create an \”exception\” block on that transaction that catches 1-to-n different exception types. For instance, you can catch \”System.Exception\” and then in that exception block, build a message detailing the error and send it out via an SMTP adapter. You can also mark the entire orchestration as long running, and then if anything fails, anywhere, it’ll trigger the compensating orchestration.