Hmm. This looks like the connection is just dropping, and then the
server times out because the client wasn't receiving anymore data. I
don't have the code in front of me, but if I remember it right the
"killed" in the end of the sshd_log line entry indicates the local SSH
client was killed, sending a signal to the server to shutdown the
task. Which it did by closing down the pipe, resulting in the write
error.
As for why the client aborted... it might be because of the SSH
KeepAlive setting. Is it set really low and the client isn't able to
get the KeepAlive ping back from the server? I don't know, you would
need to use ssh with a -v (or multiple -v options like -vvv) to see
the debug trace on the client and watch for when it sends/receives
KeepAlive packets with the server.