I’ve got a task mapping set up, but the child tasks sometimes have to retry until the external data they read is ready. When one of them enters the Retrying state, it seems to hold up the entire flow. The parent doesn’t produce any more tasks until the retrying one is complete.
Is there way around this behaviour? Ideally I’d like the parent to spawn all its tasks right away, and for Prefect to identify which ones can be run, rather than one retry blocking everyone else