SQL 2012 AAG failover unexpected

Hi All,

We have an instance of SQL 2012 running on 2 windows cluster nodes, and the instance was setup AAG with synchronous commit and automatic failover. The primary instance did failover at ~00:16 last night. I checked SQL error log, event viewer, windows cluster log,sql trace and extended events, can't find many useful info, the extended event only showed "data[hadrag] SQL server service is not alive" and "reasonSERVERDIED". 

Could anybody help me to check how to find the root cause?

Thanks

David