The agent should check if all cluster nodes are up and running, befor a update is started
We have created a backlog item to track this request.
Cluster-Aware Updating would be nice.
I want the patches to be applied after the state of the first cluster node is good again. Let the mechanism talk to the cluster-aware update component.
Vincent Bersagol commented
For SQL Always On, having a way to patch 1st the secondary node(s), then manage properly the fail over and then patch the last server