Setting the com.mendix.core.ProcessedTasksCleanupAge to such a low value does not have an immediate effect. This is because the cleanup is performed by the cluster node (or the only node if not clustered) every ClusterManagerActionInterval, which defaults to 5 minutes.
So what is the point of requiring the value to be entered in milliseconds if it can't process in less than 5 x 60 x 1000 = 300,000 milliseconds? It seems like some bright engineer could do the math for us under the covers so we can enter values that make sense, like in minutes, hours, or days.
The documentation for this is even more strange. It says the default value is "365 days for projects migrated from Mendix 9 and 7 days for new projects or projects with an empty database." So the default value is 31536000000 milliseconds for projects migrated from Mendix 9 and 604800000 for new projects or projects with an empty database.
Maybe someone could improve this?
Locally I've got the same issue with com.mendix.core.ProcessedTasksCleanupAge while the functionality for com.mendix.core.ScheduledEventsCleanupAge runs fine. I haven't tried this setting in Mendix cloud yet.
My advice: create a ticket at https://support.mendix.com