I'd like some guidance on the best practice for handling exceptions.

What we want to avoid is process instances with Error statuses that no one knows about.

In testing, I added an exception handler to a process. I then threw an exception from a server event. This worked fine; it emailed the error to me.

I then added a synchronous IPC call to a subprocess, and threw an exception in the other process. This did not percolate to the calling process; I assume I would have to add the exception handler to each process.

In doing this, though, I made an error - I didn't give permissions to create the subprocess. This resulted in the following error in the error profiles:

Type 20
Source IPCSend
[servername] does not have permissions to start the process

And this did not fire my process exception handler. How can I catch these errors?

Is the best option to add exception handlers to every process?

One alternative we came up with was to watch the k2 error databases for new entries. Has anyone tried this?

All suggestions are welcome.

Thanks much,
Bill Sorensen