I checked the message tracking logs on Exchange and nothing was ever leaving the OpsMgr server.  I tried to telnet to the smtp server and that worked.  I checked the logs and there was an innocuous message about the “Instant Notification” settings I were using, and how it was causing a notification workflow.  The app log message looked like this:

The exception text is:
Microsoft.EnterpriseManagement.HealthService.ModuleException: The specified Uri : (sip:email@im.vanity.com/home) has one or more invalid elements. A valid sip uri should be of the format sip:user@host.
Parameter name: uri —> System.ArgumentException: The specified Uri : (sip:email@im.vanity.com/home) has one or more invalid elements. A valid sip uri should be of the format sip:user@host.
 

(insert lenghty description of lots of things going crunch)

For more information, see Help and Support Center at http://go.microsoft.com/fwlink/events.asp.

So I decided it was important to revisit my IM settings, since I knew that was what I was working on the day notifications had stopped. And there it was glaring at me in the face. the last thing I changed was added “/home” to the “Return Address” under the notification settings.

Once I removed that I proceeded to get 2 days worth of back-logged alerts. It is interesting to note that an error with the IM notification workflow actually affected the Email notification also.

I’m just happy I found it and fixed it. In case anyone ever sees the Event ID 4509 on your RMS, I would suggest visiting your notification settings.

The reason for this issue is the Operations Manager certificate for this computer still has the original image build’s server name on it.

To fix this:

1. Stop the Operations Manger Health Service
2. Open up an MMC and openthe certificates snap-in for “Local Computer”
3. Navigate to Certificates -> Operations Manager -> Certificates
4. and delete the certificate with the original build’s name
5. Restart the Operations Manager Health Service.

This fix may not be the best way to solve this issue, but I have not done a thorough web search for better solutions yet.  If no one else has this documented somewhere, the steps could easily be scripted, and if I don’t find a better solution I will post the eventual script for this once I have it done.