We are still in the same boat. While rules are configured to have none but a very few systems ever go into the AH, random systems from all over the environment continue to do so. We un/reinstalled the AH function on its server, and , this time, we set up the db credentials manually ( by not checking the box that said 'use epo server's db credentials during setup) and this did not change the symptoms.
If anyone knows of a log file we can look at to see why systems are choosing the AH over the ePO , I would appreciate it!
This may be nothing but a diary-- so be it. Maybe it will help someone else.
I found the place in the agent (backup) log where one system changed over to the AH, but it just said 'trying x.x.x.x blah blah.. no reason why, no message that it had failed to contact ePO first or what have you.. still trying. I"ll check more agent logs and hopefully find something that has to do with why/how this is happening.
We opened a ticket.. they want to analyze agent.logs and orion log in debug mode. We'll see what happens.