As automations get more and more complex, there are more possibilities for failures along the way.
But this behavior is actually critical for maintaining stable running systems. We live in the real world where data is messy; despite our best efforts to get clean data all the time, invariably this will fail and something weird will come through.
Kinda like how we have a custom-timeout when using a "wait" task
If there was an "On Failure" branch at the end of the automation chain, that would allow us to define something to do in this instance (send an email, call a process, mark the record as an error record, etc.), we could catch these instances and maintain stability.
--------------------------------------------------------------------------------------------------------------------
Cheers; and as always, thanks for considering!