erlang / erjang debugger

[ToddG]

Any suggestions as to how one would get started making a better erlang/erjang debugger?

Here's what I'm thinking:

* classic debugger functionality, line and conditional breakpoints, step in step out, etc.
* system stats (avg message que lengths, messages/sec, num processes created/sec)
* integrate Kresten's awesome erlubi process viewer for drill down and process context
* REST interface, so the debugging can be decoupled from the viewer
* leverage some fantastic web dev (not it) to make the front-end really slick

I'm thinking of this project as a way to really learn erlang, too.

Other random questions:

* are there libraries like there are for malloc in C, that would allow you to monitor process creation? this way you could find hot spots in the code where processes keep dying...

* how about a similar idea for intercepting message passing? this would be useful for adding graphics to the erlubi to show data flow between processes

* suppose you had a vm that was leaking processes (memory), and was reaching some limit that would trigger a graceful vm shutdown (vs a crash)... how does one gracefully bring a vm down? is there any way to migrate work to another node in the cluster? can you migrate processes, or do people typically just push state into mnesia and shut down the vm and restart another one?

-Todd