v1.5.18 — Fix PlexNotFound retry storm in batch path

Bug Fix

PlexNotFound retry storm in batch processing — When scenes weren't yet scanned by Plex, the batch process_queue() loop retried them 3-5 times within 2 seconds instead of respecting the 30-second backoff window.

Root Cause

The batch loop was missing the _is_ready_for_retry() check that the background worker already had. SQLiteAckQueue returns re-enqueued jobs immediately because next_retry_at lives inside the job payload (opaque to the queue), so without the guard, the loop would immediately re-dequeue and retry.

Changes

• Batch loop now checks _is_ready_for_retry() before processing — nacks the job if not ready
• PlexNotFound re-enqueued scenes are marked in processed_scenes so the same batch run doesn't re-pull them