This document summarizes the process for upgrading an OpenStack deployment from Kilo to Liberty. Key steps included automating the entire upgrade process, testing upgrades using production data, improving monitoring during upgrades to catch network issues, and performing a rolling upgrade of the control nodes to minimize downtime. The control nodes were upgraded one by one by moving services off each node, upgrading packages and configurations, and validating services before bringing the next node back online. While network issues were still encountered, the process represented an improvement over the previous upgrade and helped meet requirements of less than 15 minutes of API downtime.