We do a similar thing - check several different servers for failed jobs nightly. We have one job on a master server that queries MSDB on the other servers (yes, we use linked servers). Each step in the job runs the query on a different server and appends the results to a text file. The last step is to mail the text file.
It's simplistic, but pretty easy to maintain and use. If anyone wants the code, let me know.
Betsy