What about setting the default behavior to send all of the errors to the error pipeline, where you can develop a logic to classify the errors based on the incoming data?
I have a generic error pipeline, which logs errors to a database and later picked up by a mechanism to create internal support tickets. I’d like to keep the default behavior without adding an extra pipeline. But generally, yes, i suppose that would work.
Thanks a lot - while it did require quite some refactoring (the pipelines should use accounts, not a variable as token, there is a lot of complexity reduction possible with the shared pipelines) - this is what we were looking for to get started, so i’m happy to say - after half a days worth of work - this is now in production and running our “nightly export”.
