ABRT is designed to only handle crashes on an operating-system level, i.e. it only handles packaged binaries by default and captures crashes coming from all of them. You might set up a blacklist, but that is rather tedious to get right if your only goal is to monitor the health of your own application.
sABRTooth, currently in experimental phase, was born out of the growing need to track issues in cloud applications. With only a running instance of ABRT Analytics and minimal changes in your code required, you can build basic monitoring infrastructure for your needs.
Since ABRT Analytics is not designed with this in mind, we would need improvements in the longer-term to enable easier single-use deployments. Another goal is providing support for a wider range of languages (currently only Python).
Here is a quick demo of it in action: