fix: Isolate cleanup steps in Actor __aexit__ to prevent cascading failures (#842)

- Wrap each cleanup step in `finalize()` (`event_manager.__aexit__`,
`_charging_manager_implementation.__aexit__`, `_save_actor_state`) with
independent `try/except` blocks
- Previously, if `event_manager.__aexit__()` raised, charging manager
exit and state persistence were skipped entirely
- Failures are now logged and don't prevent subsequent cleanup steps
from executing

---

Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

Author: vdusek

src/apify/_actor.py

@@ -249,11 +249,21 @@ async def finalize() -> None:
             if self._event_listeners_timeout:
                 await self.event_manager.wait_for_all_listeners_to_complete(timeout=self._event_listeners_timeout)
 
-            await self.event_manager.__aexit__(None, None, None)
-            await self._charging_manager_implementation.__aexit__(None, None, None)
+            try:
+                await self.event_manager.__aexit__(None, None, None)
+            except Exception:
+                self.log.exception('Failed to exit event manager')
+
+            try:
+                await self._charging_manager_implementation.__aexit__(None, None, None)
+            except Exception:
+                self.log.exception('Failed to exit charging manager')
 
             # Persist Actor state
-            await self._save_actor_state()
+            try:
+                await self._save_actor_state()
+            except Exception:
+                self.log.exception('Failed to save Actor state')
 
         try:
             await asyncio.wait_for(finalize(), self._cleanup_timeout.total_seconds())