Normally, the git-annex assistant makes your files be available wherever you use it, and so a copy of each file is stored in each repository. That's perfect for files you're using right now, but what about files you're not using any more?
You could just delete those files, but it's better to archive them, so you can access them later. All you need to get started archiving your old files is a USB drive, or an Amazon Glacier account.
The webapp makes it easy to make a repository on either a USB drive, or on Amazon Glacier. Once the repository is created, be sure to put it in either the small archive, or full archive repository group.
Now when you're done with a file, just move it into a directory named "archive". The assistant will notice you put it there, and next time it has the opportunity (when you plug in the USB drive, or when it can talk to Amazon Glacier over the network), will move the file's content to your archive repository.
You'll no longer be able to open the file once it's been archived. If you later want to access it, you can just copy or move it out of the archive directory, and the assistant will retrieve its content from the archive.
Note that retrieving data from Amazon Glacier takes 4 to 5 hours.