Fix _meta_run: hold lock for entire yt-dlp execution, not just scheduling

Previously the lock was released before _run(), so multiple threads could
fire yt-dlp processes simultaneously — completely defeating the rate limiter.
Now the lock is held through the subprocess call and released in finally.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

backend/services/ytdlp.py

@@ -33,8 +33,10 @@ def _meta_run(args: list[str], timeout: int = 60) -> tuple[str, str, int]:
         wait = _META_MIN_GAP - (now - _meta_last_call)
         if wait > 0:
             time.sleep(wait + random.uniform(0.5, 2.5))
-        _meta_last_call = time.monotonic()
+        try:
             return _run(args, timeout=timeout)
+        finally:
+            _meta_last_call = time.monotonic()
 
 
 def _parse_date(date_str: str | None) -> datetime | None: