Can you restart a failed Prefect run from a particular subflow/task (e.g. in the middle of the DAG)? The compelling use case is a Data Science DAG that takes more than 20-30 minutes to run and you failed in the last mile (which takes <5 mins).

No you cant. But if you use caches and storage. You can skip the computation/ io and get the saved data from the cache instead.

Thanks @Deceivious. Is there an example deployment / workflow code for using the caches and storage to skip and fetch data from cache?