Disclaimer: I don't know this is the actual problem, I just suspect it is.
I have set up an Amazon Web Services EC2 instance as a remote git server. I have had to configure it to only accept incoming traffic from a single IP address though I had wanted to make it open to all because I will be using it while I travel.
The complication here is that mobile phones cannot have a static IP address and I will be accessing the server using my android phone tethered to my laptop. This means I have to make a new rule to accept traffic from my current IP address before connecting to the server via ssh.
When setting this rule there is an option to use "My IP Address". Amazon automagically checks my current IP address and sets it in the rule. It looks like 123.123.123.123/32. The "/32" part there being the actual end of the address.
When testing the server from home I have no problem accessing the server from my desktop computer over the home network with static IP. Equally no problem accessing the server from my laptop computer over the home network. I then tethered my phone to my laptop and tried to access the server via the cellular network with the home static IP address rule still in place. As expected the connection was not made because the phone is allocated a different IP address.
BUT...then I go out for a day of coding in a nice location and try to connect to the server and it does not work. I went in to my EC2 console and set a new rule to accept traffic from my current IP address over my tethered phone using the automagic "My IP Address" setting. Then tried to connect to the server via ssh. No joy.
The connection is not made. Just times out in the same way it did when I originally tested trying to access from an IP address not specified in the rule.
So I was thinking it has something to do with ports. Is that what the "/32" above is? Or have I got that all out of whack?
What can I do to gather more information to help define what the actual problem here is?
Find an Android VPN provider that will sell you a static IP for every use. That might be a way to use a static IP as a security measure. If the VPN provider offers the service on PC as well you can set both to the same IP. Your really ambitious you could set up a VPN service on your virtual machine and then configure it to only except connections from it's self via SSH. You would have to Adjust your "security group" to except VPS connects from any where. The second method would mean that you wouldn't need two static addresses for and added layer of security.