When Observability Meets Sustainability: A Real World Experience
summary
A deployed measurement system collects ~100 metrics across four dimensions: security and compliance, reliability and availability, effectiveness of operations, and greenness and low carbon, aggregates rack and cooling sensor data plus application telemetry, and surfaces real-time dashboards from a live IBM China data center. Kepler breaks machine power down to containers including IBM Z, its agent is lightweight, and automated actions—especially AI-driven scheduling tied to real-time electricity prices—reduce PUE and lower energy and carbon footprints.
AI-written summary · about
chapters (10)
- 0:00Introduction
- 1:51Sustainability in IT
- 3:36IBM China’s sustainability framework
- 6:46Framework implementation
- 8:52Demo setup and observability tools
- 13:49Live energy monitoring demo
- 20:18Efficiency metrics and indicators
- 24:39AI energy optimization
- 27:13Carbon tracking and performance
- 29:19Q&A
related talks
33mWhen Observability Meets Sustainability: A Real World Experience
17mKepler: Powering up the efficiency of your cluster
26mAccelerating Science, Decelerating Carbon: Sustainable Computing...
40mMeasuring Energy Consumption in Clusters with Kepler
40mSustainability the Container Native Way
42meBPF-Powered Observab...
15mImportance of Observability in Platform Engineering
14mObservability simplified: Empowering engineers to make better software