I have checked arerror.log and found that there is ARERR 552 is occuring continuously.
I will suggest to check your Network setting and Firewall configuration to check why Database connection is getting lost.
Also please provide Server side SQL and Filter Logs to Investigate more on your issue.
Please be sure to mark Correct or Helpful answer if you find my inputs useful.
the following error is thrown from the OS and recorded in the arerror.log - this is due to connectivity issues between AR > DB:
[DBNETLIB][ConnectionWrite (send()).]General network error. Check your network documentation. (SQL Server 11)
You should check out the following article if you are using Virtual Machines across your environment:
The change will need to be performed on the AR Servers.
Might be a little late to post on this..
I had a similar issue. I have three AR Servers on a Database.
I first targeted on SQL error "SLQ Server 11" that seemed to be the cause, but it wasnt directly related I realized later.
Our AR servers used to log "SLQ Server 11" error once every hour in the arerror.log and the servers used to hang once in 2-3 days. Sometimes more often and sometimes not.
After observing the trend couple of times and I started looking into the "Insufficient Memory" error and contacted BMC. They helped. We had Insufficient Memory error followed by SQL Server 11 errors.
They observed that the servers were eating up 85% of the available 12G memory all the time and AR processes were merely taking 1.8G out of that. The rest of the memory was taken up by some back ground processed that I could not figure out.
I brought this to my windows server admin's notice. They said that "the VM's(WindowsServer2008R2SP1) resources are restricted at the VM". I am not sure what that meant but they did something from the Task Manager, I believe (may have killed some processes and disabled them which were not required and causing problems). But soon after they did that, the servers' memory soncumption went down from 85 to 20%. Its been 2 days now, the servers are stable and the DB error(which used to log every hour) is gone!
There is also a microsoft defect about WindoesServer2008R2SP1 memory leak. Our windows Admin might applied the a fix around that but I am not sure.
hope this helps!