Access Levels Completely Reset to No Access?

[KatherineLa]

I have a support ticket open on this already, but I'm also curious if other Admins are experiencing this behavior in their environments and if Support has been able to provide any help or insights on the cause/solutions?

Starting a few months ago, one or more of our custom Access Levels will be randomly updated from it's correct settings to "No Access" for every possible item. This causes a complete service outage for however many users are assigned that access level (and a fair amount of help tickets to me!) Today's outage occurred in the middle of a 2-hour teaching demo in which my affected users were supposed to follow along with the training. (When it rains, it pours?)

Reviewing the Audit Log each time will show no recent changes to that, or any other, Access Levels by anyone within our organization. This outage lasted 2 full hours, and with 5min left in my demo, the Access Level fixed itself, again with no changes recorded in the Audit Log by anyone in our organization. 

I don't like that something at Adobe is able to make unauthorized and completely undocumented changes to Access Levels like this. Has anyone else seen this occur? 

I've attached screenshots showing what I found while troubleshooting, what it looked like after it mysteriously fixed itself and the Audit Log showing no changes in the last 48 hours were recorded. 

[KellieGardner]

Interesting scenario that is occuring. I haven't experienced that incident in our instance.

However, I did have a ticket open for quite some time because my Group Admins couldn't see anything in the Access Level dropdown on a user. We changed absolutely nothing on our end, they just disappeared from their drop down one day. It was eventually fixed but took some time.

[Edwin_Melgar]

I have not seen this, but this would be frustrating.

We recently went through an new onboarding of a new group.  Access Levels were created by our implementation partner  and I don't see those listed in the audit log.  I wonder in kickstart import or other type of API type imports are not logged?

[VictoriaLinn]

Hi @KatherineLa - We've had this same problem occur, but maybe not as frequently as you're describing. Agree with your point above that it is almost impossible to provide proof too because the audit logs don't keep a record of the change.

What we did to provide that it was happening to Workfront Support is compare it against Preview. So, if Preview is correct, you can assume it was good during the weekend refresh, then there are no audit logs showing access was changed during the current week, so it wasn't a user in the system and something else is happening.

We found this to be happening about a year back though, wish it was fully fixed!