I'm not really sure what happened, but earlier today one of my Ubuntu boxes network (hosted on AWS) randomly died, I was able to fix it up by assigning an elastic IP and rebooting.
But since then the networking hasn't been quite right.... And I'm not 100% confident if it is an AWS issue or server issue, but my gut feel is that it is a server issue.
On the application level it will randomly work for a brief period (at normal perceived speed) or it will timeout. This happens across at least port 80 and 443. (tested using curl, wget, apt/apt-get, PHP, etc)
Currently when I run tcpdump
most of the time all the TCP responses have a length of 0, but not always, sometimes it works perfectly fine. (I can attach a copy of the dump if that will help)
I haven't made any changes to the network config either outside of the ones AWS would've done when making the elastic IP change.
NOTE: Pinging the host from the outside has been very inconsistent, once it starts pinging it's fine, but it might take 80 attempts or so for it to start working, but HTTP/HTTPS seems to be okay
I've also checked iptables and they are blank
Running Ubuntu 14.04 on AWS M3.Medium
Running as a web server and most if not all the HTTP/HTTPS inbound traffic works perfectly fine