Full-stack energy consumption test bench for developers
summary
A test bench named Argos runs a web app inside containers, collects per-container Docker stats (CPU, memory, network, disk) while a Cypress simulation replays realistic user scenarios dozens of times, converts all metrics into watt·hours using an academic model, and outputs per-layer energy and delta reports — e.g., an introduced bug causes almost 3× more network use on the device and a measurable energy increase. It’s open source, CI-ready proof of concept from the 2020 Sustainable Digital Challenge.
AI-written summary · about
chapters (13)
- 0:00Start
- 5:38Why we created Argos
- 6:45Reasoning behind the first aim
- 7:18Improvements done for the projects
- 7:51So how much did we gain?
- 8:18Using existing ecometers?
- 11:57How to measure energy consumption?
- 12:07We need to measure everything
- 12:17We need to measure it everywhere
- 12:27How to measure on the client ?
- 13:04Using docker we can
- 13:52To get realistic data we need to
- 15:33Measure frequently and automatically
related talks
23mFull-stack energy consumption test bench for developers
22mHow to (easily) evaluate the energy consumption of your software ?
13mPerformance and Energy Efficiency in Docker
08mTowards Greener Deployment: Assessing Energy Efficiency i...
28mMore Data Please: Hands on Green Cloud Experiments
31mEvaluating the Energy Footprint of GitOps Architectures: A Benchmark Analysis
25mGreen Benchmarking of BEAM Languages
26mTales from the QA crypt