Getting the download count of a specific S3 object

I've got a number of S3 objects that are available to my customers. Since I'd like to bill my customers by usage, I wondered if there is any smart kind of way to get the number of times a given file has been downloaded?

Alternatively, I suppose I could parse the log files provided by S3, but with 10m+ fetches per customer this might be bit of a task.

Any ideas?开发者_StackOverflow中文版

If you let them download via HTTP then you could use a URL shortener service to redirect to your file (like goo.gl). I suppose they can handle load and they also provide some nice stats next to total clicks.

You could redirect through a script on your site that increments a count before redirecting to the S3 URL.

There'd be some assembly required and it needs some cleanup, but I wrote an app for this based on node.js and Couchbase. I need to update it a bit more, but it will count HTTP 200s based on object size rather than just hitcount alone (which is inaccurate) and I've just been updating it to aggregate HTTP 206s.