Terminating connection due to serverless scale event timeout

Question

I am getting sporadic connection timeouts when running on Aurora withe the following message:  
  
"terminating connection due to serverless scale event timeout"  
  
This happens when increased load on the server causes the cluster to scale.  
  
Has anyone seen this on their cluster and have any advice on how to resolve it?  
  
Cluster Details  
=========  
  
Engine  
Aurora PostgreSQL  
  
Region & AZ  
us-west-2  
  
Capacity settings  
Minimum Aurora capacity unit  
2 capacity units  
Maximum Aurora capacity unit  
384 capacity units  
Pause compute capacity after consecutive minutes of inactivity  
5 minutes  
Force scaling the capacity to the specified values when the timeout is reached  
Enabled  
  
Encryption  
Enabled  
  
Maintenance window  
sat:12:30-sat:13:00 UTC (GMT)

Accepted Answer

Looks to me like this is intended behavior: you have enabled "Force scaling the capacity to the specified values when the timeout is reached", which means that connections which block a scaling event due to active (probably long-running) transactions will be terminated when the scaling timeout is reached. If you want the scaling event to wait for all active transactions to finish, you'll need to disable that setting...  
  
-Kevin J (PM for Aurora PostgreSQL)

Answer

Hi, I am currently having the very same issue.  
  
But at least for me, if I disable the option to "force autoscale", I get an error in the Aurora events log with the message: "Scaling point wasn't found".  
How to proceed?

Terminating connection due to serverless scale event timeout

Cluster Details

相关内容