Flaky test: tests/sentry/feedback/lib/test_label_query.py::TestLabelQuery::test_query_recent_feedbacks_with_ai_labels

Failing test: `tests/sentry/feedback/lib/test_label_query.py::TestLabelQuery::test_query_recent_feedbacks_with_ai_labels`
Sentry sha: ab471d2d1ede60c8a8c1e3911b78a71a83261eae
Run: https://github.com/getsentry/sentry/actions/runs/22271634857

```
tests/sentry/feedback/lib/test_label_query.py:140: in test_query_recent_feedbacks_with_ai_labels
assert result[0] == {
E AssertionError: assert {'feedback': ..., 'Security']} == {'feedback': ...Performance']}
E 
E Differing items:
E {'labels': ['Authentication', 'Security']} != {'labels': ['Performance']}
E {'feedback': 'b'} != {'feedback': 'The app crashes frequently when loading data'}
E 
E Full diff:
E {
E - 'feedback': 'The app crashes frequently when loading data',
E + 'feedback': 'b',
E 'labels': [
E - 'Performance',
E + 'Authentication',
E + 'Security',
E ],
E }
```

Run tests/sentry/feedback/lib/test_label_query.py::TestLabelQuery::test_query_recent_feedbacks_with_ai_labels at Sentry SHA ab471d2d1ede60c8a8c1e3911b78a71a83261eae and read the assertion and test setup around line 140. Trace why the returned feedback and labels differ between runs; done means the test consistently produces its intended result without masking the failure.

Written by the indexing model from the issue text.