I'm working on an application that'll be utilizing the rclone's RPC interface, though I want a way to persist _async jobs past the 1-minute mark that rclone sets by default. It's mainly a concern for when my application freezes past the 1-minute mark due to potential future bugs, and I just want to make sure my program can gracefully handle poor situations.
My initial thoughts for such would be allowing passing a flag such as _async_keep: true (or similar), and then having an endpoint such as job/clear that accepts a job ID, which would clear the job from whatever rclone stores it with.
Any thoughts? I'd be more than open to getting the PR for this going.