Assuming you have millions of job a days that runs on your hangfire instances (and that quickly become true when you start using batches) retention time longer than a day could take significant amount of space on your Redis instance.
Moving jobs executions details on slow and cheaper storages like RDBMS or plain filesystem would free retention policy to fill up expensive RAM on Redis.
The RedisWriteOnlyTransaction.ExpireJob method seems to be a good extension point for that approach.
Assuming you have millions of job a days that runs on your hangfire instances (and that quickly become true when you start using batches) retention time longer than a day could take significant amount of space on your Redis instance.
Moving jobs executions details on slow and cheaper storages like RDBMS or plain filesystem would free retention policy to fill up expensive RAM on Redis.
The RedisWriteOnlyTransaction.ExpireJob method seems to be a good extension point for that approach.