Now that we have an understanding of what’s going wrong, we can also rule out Azure as criterium. This should misbehave in all environments with windows push-agents.
Now the big question: When will it be fixed? Will it be sooner if we send a MR/PR?
In my understanding, we need to change our_systemtime that is mentioned here
I think the “easiest way” would be a support ticket referencing this thread and Andreas’ debugging findings.
The problem is known and conceptual due to the push agent and how we process its data. We have not found a good and easy solution yet. At least with werk #17319 we are documenting it clearly.
We created a new timesync check for Windows, which addresses the issues with the System Time service.
Please give it a try, and share your feedback with @MSaleh