I'd say you are the second 🙂

I'm setting up an Always On Group these weeks.

I also have some large tables that are partitioned by month. The database ist part of the Always On Group.

My first test of switching out an partition into another table worked fine at the first try. The partition was empty but the whole process worked.

The next time we have a partition full with data. I do not expect any problems as the mechanism of the sliding window seems to work. The interesting thing is: We use synchonous mode. How long will the transaction of switching out the table really need and what does this mean for the primary replica? How long will it be "non writeable". As our network connection between the primary and the only secondary replica is quite strong I do not expect any problems here neither.

In the end this means for me: As the sliding window itself works great it seems to be like on any "non-always-on-database", only to keep in mind the any transaction has to be committed from the secondary replica when using synchronous data transfer.