In the _remove routine when client.end is called we are seeing frequent throws because the socket has already been closed. We wrapped the call to client.end in a try catch and everything seems to be working. Looks like RDS (AWS postgresql implementation) may be closing the socket on the other end. (This is happening in the idle timeout handling not through any explicit call by our code.)
In the _remove routine when client.end is called we are seeing frequent throws because the socket has already been closed. We wrapped the call to client.end in a try catch and everything seems to be working. Looks like RDS (AWS postgresql implementation) may be closing the socket on the other end. (This is happening in the idle timeout handling not through any explicit call by our code.)