The water industry is often and unfairly criticised for being relatively slow to adopt IT as part of its operating processes. To address this, the FIWARE and Fiware4Water projects were tasked with developing open-source data definition, management, and processing, and applying it to the water sector. In this work, we are reviewing our experiences with developing FIWARE-based solutions for the aqua3s[1] and Fiware4Water[2] projects across the whole project lifecycle, from our initial of installing and learning to work with FIWARE context brokers[3], through building and maintaining synthetic digital twins for load and edge-case testing development and debugging and finally looking at longer-term data management and archiving. Through our work on these projects, we have started to develop a wealth of best-practice, strategic choices, and approaches to be avoided for successful and long-term context broker usage.