Details
-
Bug
-
Status: Open
-
Major
-
Resolution: Unresolved
-
None
-
None
-
None
Description
It is possible to set a maximum parallel of sling jobs on a queue configuration, but in systems where the platform itself already makes 20 queue's and the customer's code also add's their own queue's, this configuration makes no sense because it is set per queue and not overall. The sum of the parallel could be way more than the amount of cores available, making the system unworkable if the separate queue's all have jobs at the same time
It would be logical that you can also set a maximum parallel overall for all queue's combined. Is there a reason that this is not already the case? Would there be anything against building this in?