Most people working with the ESB Guidance from Microsoft will likely *never* run into this (we sure didn’t during testing :)) , however, as it could lead to some confusion, I thought I’d do a mini-post about it.
At a client site, I was creating and publishing Fault messages no problem. Then, I moved it into a scope’s exception handler, and it started failing, telling me that I had exceeded the 256 char limit on promoted props.
After some digging (time passed, this was a good challenge!), I discovered that the root cause was that in the ESB Guidance exception management sub-system, we are auto-populating the FaultMessage.Description property. What happened when I moved it into my scope was that it was populating it with the rather verbose SOAP fault message which includes callstack, etc, and *that* was exceeding the 256 char limit.
Solution: In the Microsoft.BizTalk.ESB.Exception.Schemas project, I disabled the promotion of the “description” property in the FaultMessage schema, and all works fine. This is a reasonable fix, as this is a system-populated field, and I think it’s unlikely you’d set a filter on this system-populated exception info. If you wanted to do something like that, you’d more likely set it on the developer-supplied “FaultDescription” property (in my opinion).