I assume that with embedded applications you are referring to mashups (as opposed to oem)? We don't provide any scalability tools for this, but you could layer your testing. In the underlying app providing the visualizations to your page, create a sheet with the visualizations as they are presented on your page. Then run a test with the scalability tools towards these sheets. Then test rest of your page as you would normally using e.g. JMeter. Technically you could write JMeter tests for the sense visualizations on your page too, but the websocket support in JMeter is not very good.