This message was deleted.
# general
s
This message was deleted.
g
with sutiable intermediate handoff config, the main remaining reason to not run them indefinitely is that it interferes with rolling updates if you do them the way described here: https://apachedruidworkspace.slack.com/archives/C0303FDCZEZ/p1661309608172939?thread_ts=1661309249.992659&cid=C0303FDCZEZ & also the task log files get too long 🙂
👍 1
we usually use durations of 1h – 4h in production
c
another reason is that if a task fails, you will be missing all of the data from the beginning of the task and it will have to start over!