Hi T,
We have some processes where we throw an exception within a block where the block has recover-resume actions. This obviously will not terminate anything but allow the process to continue after handling the error according to the path you defined.
From my understanding and experience, you should do as little as possible between a recover and a resume to avoid errors bubbling up, coming back on a retry and generally cause looping. To my knowledge, the only thing you can safely do between a recover and a resume is collecting the exception data, such as details and type. This fits right in on the statement you quoted '...avoid using stages that may throw additional errors...'.
------------------------------
Happy coding!
Paul
Sweden
------------------------------
Happy coding!
Paul, Sweden
(By all means, do not mark this as the best answer!)