Is there an option to have "auto" cleanup of persisted jobs? in particular when using the Redis storage backend we eventually run out of memory. Not such a problem for other backends, but I think if a solution was implemented it would be part of the core traits