Migrated my 5.8.4 Docker container (mbentley) to 5.9.9.  It's a bit different now, you have to be at the Global level to do a controller backup.

I spun up the new Image and applied the backup file I had just made (1.1MB).  It had approximately 6mon of data from at most 5 network devices but it took a full 30 MINUTES to restore, during which time the controller was unavailable!  CPU was nowhere near pinned (4-5% typical).

I think you *REALLY* need to split the config data from the history data in the .cfg file, and then background the 'restore' process for the historical info.  Emphasis should be on getting the controller back online ASAP.