Why does the 1C Enterprise server periodically fall off?

Question

Debian

Good day friends, I ran into a problem, I can not figure out how to solve it. About everything in order.
Not so long ago, we started the transition to 1C BP 3.0 and everything that is based on it, compared to 2.0, it began to slow down 5 times more, bases opened from 2 to 5 minutes! The solution came quickly, MS SQL Server.
Since this is the first experience of setting it up for me, I started training in the Hyper-V virtual machine, everything worked more or less until I loaded the database of the agricultural department there, after about 2 - 3 hours of work, the server stops working, the database does not connect, even directly on the same server. On the local network, the error "1541 descr = server is not available (does not respond, crashes or the port is occupied by another application"). First I sinned on the network card, then on the firewall (disabling it also did not help), and yesterday I took a physical server , set everything up, installed it, transferred the databases and just an hour ago the same problem!All drivers are updated, the server capacity should be enough (Xeon X5606 2.13 Ghz/24Gb RAM DDR3 1333/LAN 1Gb) the network topology is "Star". the server is not included in the Active Directory domain,
Briefly about the software:
1C Enterprise 8.3.8.2088
MS SQL Server 2014 SP1
Windows Server 2008R2
I would be glad for any advice.
Update:
A similar problem appears on Debian. With one base UT10 works fine. When migrating an old database (accounting) from a file version to SQL (postgres), periodically once every 1-2 weeks the databases become inaccessible until the srv1cv83 service is restarted. There is a feeling that the more databases are transferred, the shorter the period of work before restarting, so when transferring 5 databases, the period of work was 2-3 days.
One UT10 database is accessed locally (on one machine), when other databases are added, work with them begins through clients over the network, but at the time of a failure, the databases are not available in any form until the srv1cv83 service is restarted.
In the technological log, only the following can be distinguished from the suspicious one:

52:20.692004-0,EXCP,0,process=rmngr,p:processName=RegMngrCntxt,p:processName=ServerJobExecutorContext,OSThread=1872,Exception=81029657-3fe6-4cd6-80c0-36de78fe6657,Descr='./src/ServerJobScheduler.cpp(1924):
81029657-3fe6-4cd6-80c0-36de78fe6657:  server_addr=tcp://1cserver:1560 descr=Сервер недоступен (Не отвечает, завершается аварийно или порт занят другим приложением) line=1089 file=./src/DataExchangeTcpClientImpl.cpp'
     52:27.139000-0,EXCP,2,process=rmngr,OSThread=1861,DumpFile=Received signal 15. Terminatingo

36:11.720015-0,EXCP,2,process=rmngr,p:processName=RegMngrCntxt,p:processName=ServerJobExecutorContext,OSThread=1892,t:clientID=2841,t:applicationName=ServerProcess,t:computerName=1cserver,Exception=81029657-3fe6-4cd6-80c0-36de78fe6657,Descr='./src/RemoteInterfaceImpl.cpp(810):
81029657-3fe6-4cd6-80c0-36de78fe6657:  server_addr=tcp://1cserver:1560 descr=recv returns zero, disconnected line=1587 file=./src/DataExchangeTcpClientImpl.cpp'

57:47.923000-0,EXCP,1,process=rmngr,p:processName=RegMngrCntxt,p:processName=ServerJobExecutorContext,OSThread=1864,ClientID=2868,Exception=NetDataExchangeException,Descr='Нет ответа от сервера server_addr=tcp://1cserver:1560 timeout=30000 line=2232 file=./src/DataExchangeTcpClientImpl.cpp'
57:47.923001-0,EXCPCNTX,0,ClientComputerName=,ServerComputerName=,UserName=,ConnectString=

Answer 1

Play around with the 1C cluster settings.

Answer 2

Logs look both screw and 1C.