Log monitor state or a custom message when it is in a hot state at the end of a schedule

[nvartolomei]

Liveness specification fails if a monitor is in a hot state at the end of a schedule. I have some monitors which contain state about send requests and handled ones. It would be much easier to debug the cause if this state information could be dumped at the end of a schedule. Either implicitly or via a custom "hook" in the machine.

Printing state on every event is too verbose and has high overhead.