Upgrade HA-cluster - Documentation feedback: /docs/sources/setup-grafana/set-up-for-high-availability.md

Hi, We had a discussion in our team regarding grafana upgrades on a HA-Cluster. We have two cluster nodes using the same garafana db. If we upgrade the nodes one after another, so at first we upgrade node 1 and leave node 2 running, we have HA while upgrading. But there might be a change in the db schema and we'd have a new-version-node (1) which would update the db schema on start, while the other old-version-node (2) is still running and using that db. That might potentially confuse the old-version-node (and, if the nodes start to fight, somehow, also the new-version-node). This might not be a problem if you take precautions so that an upgrade of an HA-cluster without downtime is always possible, but I don't find anything about that in the docs. Neither if it is possible nor why we shouldn't worry about the db being corrupted. I'd like the docs shedding light on that. Also, if upgrading one node while the other one keeps running is *not* recommended, a note about that would be helpful, too. Just to plan the upgrades and have documentation to point to. Thank you