Thanks everyone for the input on this thread. I don't see how MS can expect anyone to use this approach for deployment in a large data warehouse. If you have multiple jobs using the same packages running at the same time, how will you know which errors were associated to which jobs? If they are running at the same time, you will need to dig through a ton of information to get there. In a large IT shop, I can see someone spending a half hour on one failure to gather information that used to take 30 seconds. Has anyone used this approach on a large scale?