Release v1.23.5

Fixed

• A finished job no longer stays running in storage when clearing or writing its final state fails during a storage outage. The job now writes its final status before clearing its state, and keeps retrying that write (from 1 second up to 60 seconds) until it succeeds. If the clear fails, the job is still recorded as finished and retention cleanup removes the leftover record. A finished job is not completed a second time after a restart. (#100)