I agree. However some things should be stated first:
1.) Without modifing the web.config and turning debug="true" in the scheduling.provider, the log doesn't provide much information.
2.) Even after doing so, it still doesn't give much.
In reviewing the code in Processgroup.vb, which is where I think I am having issues, there is no logic to handle a failure when creating the instance of new event class. If you look at the method "Run", there are several Try/Catch constructs, but none of them catch the actual failure if process=nothing???? In other words, the Catch routine only does anything if PROCESS=something. Throwing an event out to the log if PRocess=nothing would be a real help, since I think it is having issues there.
I did forget to remove the default workspace from my scheduler client application, which is what I thought the cause might be. However, this did not correct it. I know my scheduler client is not even making into the NEW of the class because of debugging logic I have placed in there. So, I am 95% sure I am not getting a type instanced in the RUN routine.
I have tried to place code in there to correct this, but everytime I attempt to compile that module, all "heck" breaks loose and I cannot get it run again.
I have asked for this process to become a little better since 3.x. DNN 2.x at least went out and built a list of what was valid, so that you did not have to type in the class/assembly name. I never really got a straight answer as to why that was removed, but it would be nice to have that back. It would also be nice to have a failure event logged when the PROCESS object cannot be created.