Large photo uploads no longer trigger analyzer cascade

Fixed redundant self-heal logic that was causing orchestrator pile-up during batch uploads.

Bug Fixes

During large photo uploads, the contents phase was dispatching the analyzer orchestrator on a 15-second cooldown from the browser poll. Since 15 seconds exceeds the orchestrator's debounce window, every poll tick spawned a fresh analyze-work-area run—stacking 40+ runs during a single upload. The fix removes this redundant frontend self-heal: the server-side analyze-watchdog task (running every 2 minutes) already owns re-dispatching stale analyses when no page is open. With the frontend dispatch removed, uploads now proceed without unnecessary analyzer restarts.