Before I get directly to your question, some background is required.
There is a very popular concept called DevOps, which you may have heard of. The idea is that developers and sysadmins should work closely together throughout the lifecycle of a product in order to reduce friction, increase efficiency, and maybe save some people's sanity. When this is done, even halfway, it works very well and delivers results.
Unfortunately there is a completely different definition of DevOps. In this misconception, there are no sysadmins and developers attempt to do operations themselves. This can work for a while, especially when you have a service like AWS maintaining infrastructure for you, but even virtual infrastructure needs operations, and Dev and Ops skillsets don't really overlap that much. So this usually ends in disaster. It has resulted in popular online services shuttering entirely, and in one case many of you have heard of, a developer going to prison because he did not have sufficient ops knowledge or experience to secure his (illegal) web site.
(If your company is following that second definition, get help now before it's too late.)
So. The ways in which developers and sysadmins use a service like AWS frequently differ. (Similar issues arise with Docker, which I've written about on mSO.) Here is a sample of the things that developers and sysadmins may do with AWS:
Developers
- Write code for new features and bug fixes
- Check code into a repository
- Run unit and integration tests
- (In some cases) Deploy code
- (In some cases) Provide the environment (CloudFormation, AMIs, server OS and software selection) in which code runs
Sysadmins
- (In some cases) Deploy code
- (In some cases) Provide the environment (CloudFormation, AMIs, server OS and software selection) in which code runs
- Monitor servers for problems and fix problems that arise
- Maintain a secure and up to date environment
While today there is some overlap, traditionally the process of deployment and defining and providing the runtime environment have been ops tasks. For better or worse, developers are sometimes doing these tasks, often without the necessary perspective, training or experience. (Which has in part led to the popularity of Ubuntu, but that's another discussion altogether...)
For instance, in one case sysadmins may maintain the server on which Jenkins runs while developers manage Jenkins itself (which in my experience usually works best); in another sysadmins may manage both; and in a third case developers may manage both.
Ultimately, while there is always going to be topic overlap (and as I've said before, topic overlap is OK), most questions are going to be obviously development questions or obviously operations questions, and should be asked on Stack Overflow or Server Fault respectively.
One possible rule of thumb is that if the task needs root/sudo, then it is either most likely an ops task, or the environment isn't set up correctly (in which case you probably forgot to hire sysadmins).
For those that fall into the middle gray area...
make your best guess and bring your hazmat suits.
With all that out of the way, I don't think it makes sense to have a single Stack Exchange site for "cloud providers" as it would have two mostly distinct classes of questions, and readers of one class would not be interested in the other class.
Finally, it should be noted that many of the questions on SO tagged with AWS actually have nothing to do with AWS at all; it is only incidental that the person asking the question is an AWS user. When you see these, I would recommend removing that tag.
