Another thing to consider is writing a separate script that lives on your origin web servers that is a wrapper for a number of tests performed on the application. I’ve done this in the past so that the BIG-IP health monitor can be as simple as possible while leaving the logic in a script on the webservers. You can also build some caching into the test script as well to prevent overloading your backend with excessive requests from health monitors.
For now i´ll try the EAV monitor. I´m not so familiar with server scripts.
In this example (DevCentral - An F5 Technical Community), if i put my request at the URI and the string i expect at the RECV, will it work without the 5,120 bytes limit?