2 replies
BP
Hmmn... my post didn't seem to appear. Try again :-)
often we have 'long - running' tasks in the queue. In order to stop the next queued up task failing we have to extend the timeout parameter.
I would like to have a halfway point between a task sitting endlessly in the queue and not being claimed (e.g. if i extend the timeout to a huge number) or that task aborting due to not being claimed because of a long running task.
It would be good to keep the current max timeout limit, but also have a midway timer that can send a notification to project admin that task has not been claimed for a time. Project admin could either bump the item up the priority list, or take whatever action deemed correct BEFORE it aborts the task
Brooke,
its not quite as “dumb” as it seems- at least it should not be? We’ve left the name of the setting the same, but shifted the behavior to: how long does the task sit in the queue at the same position without being claimed? I don’t recall how many task servers you have deployed, but this makes it less likely to be an issue if there are many tasks backed up, but the whole bunch is inching forward. I suppose though if there’s one long running task, and only one task server to do the second task in the queue, it could be problematic…
but we’ll chew on your idea…