I would use the Quartz scheduler for this.
I have used it successfully in the past, and apparently it has a cluster mode (which I haven't tried). It performs load balancing and can use any JDBC database for coordination.
It is open source, written in Java.
I don't know of a ready made solution for this problem (and no fancy name) but would rather set it up myself (in Java). Also my JMS skills are "in development" so there might be better solutions, that combine parts 1&2. And I'm not quite sure if I understood your problem correctly.
I assume the workers handle their Database connections themselves so I ...
HTTP Client Libraries
Usually we do that kind of successive-URL-call work directly in our own code, using HTTP client libraries to assist.
HTTP client libraries such as Apache HttpComponents Google HTTP Client Library, and jsoup allow you to set a timeout, and they throw exceptions when various problems arise such as a 404.
So you write a series of ...
I would suggest that you take a look at python and pandas:
Python has a number of tools available for extracting meta data from audio files, assuming that is what you are going to cluster on, e.g. phHachoir.
Pandas provides a lot of data visualization frameworks, and can work with more, including Bokeh & plotly for web interfaces, there are even tools ...