We'd like to run the service described in bug 1163698 on an AWS VM. I think a t2.small instance running Linux will be suitable.

is this a vm to use during development, or is it a production? what level of criticality is this service? if it's anything beyond "i'm not fussed if it's down for days at a time" then i'd suggest approaching mozilla infra for deployment options. what external connectivity does this system need (inbound and outbound)?

A VM to use for development would be beneficial now, but it sounds like we should talk to mozilla infra soon. This wont be useful if it goes down, but nothing beyond that will break if it does. For connectivity we need to receive pulse messages and post to self-serve.

We're going to use Heroku for this.