The Deutsche Bahn's Path to digital sustainability — Gualter Baptista
summary
Deutsche Bahn frames developers as key actors for carbon reduction and proves a platform route: Kubernetes + platform engineering surface real‑time energy and CO2 estimates using Kepler, eBPF-to-Prometheus metrics and CloudCarbonFootprint/electricityMaps. Tested on ~900 namespaces and ~10,000 users, measurements show 70% average CPU but only 15% app usage and a modeled upside of 49% CPU and 38% RAM savings by autoscaling/downscaling. Kepler deploys on OpenShift in ~20 person‑days, runs with minimal overhead, and drives actionable GreenOps that also cut cloud costs.
AI-written summary · about
chapters (14)
- 0:00Introduction and motivation
- 4:26Climate urgency and Deutsche Bahn
- 9:06GreenOps and developer motivation
- 13:25Challenges in measuring digital emissions
- 18:57Platform engineering role in GreenOps
- 23:39Data models and real-time metrics
- 27:35Current gaps and future steps
- 29:23Q&A: Hardest part of GreenOps
- 30:09Q&A: ARM vs AMD64 footprint
- 31:11Q&A: Deutsche Bahn’s digital CO2 emissions
- 31:52Q&A: Cost and emissions correlation
- 32:37Q&A: Public administration support
- 34:02Q&A: Metrics overhead cost
- 35:07Q&A: Tool integration time
related talks
12mKeynote: Cloud Native Sustainability - How You Can Help Right Now!
59mHow Kubernetes optimisation can combat climate change
31mBridging Product and Engineering for a Greener Future: Embedding Sustainability into Software Teams
23mDriving cost-effective solutions to reduce carbon footprints
41mToward a carbon-aware cloud
46mMaking Software & Data Architectures Sustainable
35mSustainability Research the Cloud Native Way
20mEnergy Efficient Placement of Edge Workloads