When Observability Meets Sustainability: A Real World Experience

Abstract

Sustainability has gained significant attention due to the increasing concerns around climate change and energy scarcity. What do we mean when we talk sustainability in the context of computing? How do we measure and monitor the sustainability of our applications, platforms, infrastructures and facilities and then identify opportunities to improve it? In this talk, we will share a comprehensive measurement system of Sustainability in the computing with around 100 metrics on 4 dimensions - “security and compliance”, “reliability and availability”, “effectiveness of operations”, “greenness and low carbon” and requirements to the current observability systems to collect, visualize, analyze and optimize these quantitative measurements. We’ll then share our practices in a real-world data center to monitor these metrics from full-stack of the computing, analyze and optimize improvement opportunities, and then automate actions to continuously improve these Sustainability metrics without compromising performance and availability of our systems Finally, we’ll give a live demo of the whole system in our data center and show real-time dashboards of these sustainability metrics, how we analyze them to identify improvement opportunities, and take actions to optimize these metrics.