Thanks for bringing up, we'll work on making it more clear. Installing a patch temporarily stops the Portal service, which may trigger a fail over event, so it may come down to how much downtime is acceptable. One approach is to stop the standby first, apply the patch to the primary, apply the patch to the standby, and then start the standby if the install process doesn't start it automatically. I suggest you be in charge of making sure there's no failover event; depending on the speed of the install of the patch, the original primary may come up before the original standby had a chance to promote itself to primary which may cause problems.
