New Response

« Return to the main article

You are replying to:

  1. This is an easy trick, and I'm sure that I'm not the only one to think of it. However, it is so handy, that I thought I'd share.

    We've got many applications that send automatic e-mails. Some are Java, some LotusScript, etc. We were having problems with these messages not going through (delivery failures) and no one getting notified. This was especially true when the e-mail came from the server.

    Our first solution was set the from field (using Principal for LS) to a real e-mail address, so that delivery failure notifications were delivered to a real person. This wasn't the best solution as the person would eventually quit or worse yet, be inundated by messages that they didn't care about and ignore them.

    The second solution was to create a mail-in database for webmaster@domain.com. We then made sure that all automated mail came from webmaster@domain.com. Then we added one line to the bottom of each message. Program Name: x.nsf. An hourly agent parses through the body of each message and searches for this string and sets a hidden field with the program name. Then, a view is set up that shows all messages categorized by program name and then categorized by message type (delivery failure, reply, etc).

    Now when I'm on-call, I can monitor only the mail that comes back from my team?s programs and ignore the others.

    A nightly agent cleans up mail that is over a week old so that the database doesn't fill up.

Your Comments