Ali recently encountered a strange cloud burst RDS database of nearly 100% CPU problem, met three times.
The first was carried out on October 12 Ri (Zhouliu) 3:24 in the morning when the load is extremely low began, found the morning after the switchover, returned to normal.
The second also appear in the October 12, will be shortly after the switchover, the switchover was found later, before switching back to server problems, returned to normal.
The third time was yesterday (October 18) 23:15 began to appear, after this morning (also Saturday) found that the main switch back to normal again prepared.
Ali cloud RDS model using a SQL Server 2016 Standard Edition, is a September upgrade from SQL Server 2008 R2 up, configuration is a 16-core CPU, 32G of memory, compared with the previous configuration before upgrading to the next level.
Problem is very strange, yet to find clues, remains to be seen.
We have not submitted a ticket to Ali cloud, because before each encounter similar problems, Ali cloud is to let us optimize SQL, but no one was resolved by optimizing the SQL.
Sorry, this sudden cloud heart disease on the core issues to bring you the trouble, please understand.