Thanks @eskimo for suggestion. We checked with third party crash reporter and found that there was an issue in backend and is fixed now.