The rules have changed. The shift from static to dynamic infrastructure requires a change in approach to monitoring, from host-based to functional role-based. Connectivity moves from remote polling to publish-subscribe and push APIs. The control plane moves from point-and-click interfaces to infrastructure-as-code workflows and self-service, developer-friendly APIs. Now — more than ever before — organizations are faced with a deluge of data in various formats. As operators and as natural system integrators, our response to these challenges is likely to roll our own solution, construct a scalable data collection and processing pipeline and combine a number of best-of-breed tools. This is an idea and journey that I am all too familiar with. In 2010, as an operator, working for an early Cloud adopter, I was faced with these challenges, and fought for better visibility. I hacked for the greater good. In this talk, I will give an overview of what I consider to be attributes of an effective monitoring pipeline. I will recount my experience in creating Sensu, the open source monitoring pipeline, and the pitfalls the project has encountered. I will then live demo Sensu monitoring pipelines in action and make my case for not rolling your own solution from scratch.