fix: Isolate tracking failures from judge evaluation results

Wrap tracker.track_judge_result() in try/except inside _run_and_track
so a tracking failure (e.g., shutdown LD client) does not propagate
through the wrapper task and destroy the evaluation results that were
successfully computed. Restores the resilience of the previous
add_done_callback approach, which asyncio handled with isolation.

Author: jsonbailey

packages/sdk/server-ai/src/ldai/managed_model.py

@@ -68,7 +68,10 @@ async def _run_and_track(eval_task: asyncio.Task) -> List[JudgeResult]:
             results = await eval_task
             for r in results:
                 if r.success:
-                    tracker.track_judge_result(r)
+                    try:
+                        tracker.track_judge_result(r)
+                    except Exception:
+                        pass
             return results
 
         return asyncio.create_task(_run_and_track(evaluator_task))