After having to babysit a few jobs to get them to complete, an idea of submitting a way to control retries from a policy and or client level and how long it should wait for that retry to override the global retries setting. this would help with a few peskey error 41s that are not avoidable.