Postgres server running on EC2 closing every once in while. Unable to isolate the issue here - postgresql

Question

I have a database connection configured on my Django application to a Postgres server.
This server is running on an EC2 instance.
Every once in a while (3 times as of this post) I see this when I run my django App
Is the server running on host "xx.com" (13.xx.xx.xx) and accepting
TCP/IP connections on port 5432?
When I run this : sudo service postgresql status
I get
postmaster dead but pid file exists
and when I run sudo service postgresql restart I get
Stopping postgresql service: [FAILED]
Starting postgresql service: [ OK ]
I tried deleting postmaster.pid file in /var/lib/pgsql9/data
But that does not seem to solve the problem.
It seems to work fine when I restart the server. But that is Windows-user type solution I am using.
-- I am still new to using postgres.
Here is what the log file (in /var/lib/pgsql9/data/pg_log) had when the server went down:
sh: line 46: kill: (5102) - No such process
sh: line 46: kill: (5103) - No such process
sh: line 46: kill: (5105) - No such process
sh: line 46: kill: (5106) - No such process
sh: line 46: kill: (5126) - No such process
sh: line 46: kill: (5127) - No such process
sh: line 46: kill: (5130) - No such process
sh: line 46: kill: (5150) - No such process
sh: line 46: kill: (5151) - No such process
sh: line 46: kill: (5154) - No such process

Answer 1

Your application is clearly not able to reach to the server.
The error states that the application/client is trying to find a server on xx.com where it cannot find the postmaster service running on port 5432 and accepting connections.
From the other information you provided, it looks like there's something killing the postmaster at backend and when this happens the postmaster.pid file is not removed. And as long as there's a postmaster.pid file, the service status will show this message postmaster dead but pid file exists.
Try looking into the database logs (/pg_log) for more information on the crash part.