Can we please add:
-We recommend a minimum of 3 nodes in any production deployment. This is so that auto-failover can be enabled, that you don't have a single point of failure should one node go down, if one node does go down the load of restoring that node is shared by 2 instead of 1, and so that the need for growth can be spread by 3 servers instead of just two and online upgrades can be performed without bringing the cluster down to one node.
-We recommend (and in fact enforce) no more than 10 buckets per cluster due to resource requirements.
-A link to the View writing best practices
-We recommend separating the disk paths of config files, data files and index files.
This would also be a great page to have a set of "added" reading at the top:
-Perry's rebalance blog
-Perry's sizing blog
-Link to our "learn" section for more content