The reduced impact should not be used as an excuse for pushing higher risk
updates, only updates that could be considered appropriate to push to all users
should be A/B tested.
This doesn't make sense. A major reason that you push to a fraction of your users is that you think that the update has risks and you want to contain them by doing a slower rollout. Otherwise, you would just ship to everyone. Perhaps the problem here is around the definition of "risk". It's one thing to expose users to the risk that their software won't work well and another to expose them to (for instance) the risk of compromise.
Likewise, not pushing the new behaviour to any user
should be considered appropriate if some users are to remain with the old
behavior.
S 2.1.3 reads:
This doesn't make sense. A major reason that you push to a fraction of your users is that you think that the update has risks and you want to contain them by doing a slower rollout. Otherwise, you would just ship to everyone. Perhaps the problem here is around the definition of "risk". It's one thing to expose users to the risk that their software won't work well and another to expose them to (for instance) the risk of compromise.
I don't understand what this is supposed to say.