Effective and Efficient: Learn from AGL’s Delive...
summary
AGL’s delivery pipeline is measured end-to-end with simple, repeatable metrics and an open ETL stack, showing that builds typically flow in lead time 19–43 hours across distributed hardware testing and that change failure rate is rising over multi-year history. Data comes from Gerrit, Git repos and uses Grimoire Lab and Jenkins to produce dashboards per stage (commit, integration, validation, verification). Analysis finds different behavior between meta-agl and meta-agl-demo, exposes recent exponential lead-time spikes, and surfaces hypotheses to investigate (hardware, platform changes, staffing).
AI-written summary · about
chapters (15)
- 0:00Data-driven software production
- 3:57Business intelligence in automotive
- 7:28Nine-step data analysis
- 10:39AGL’s delivery process
- 14:36Data-driven decision risks
- 17:25Key delivery metrics
- 19:15Open-source data tools
- 20:24Data visualizations
- 22:51Delivery process behavior
- 27:43Detecting delivery instability
- 31:26Change failure trends
- 33:35Lead time correlation
- 37:07AGL’s data insights
- 40:18Cloud for embedded systems
- 41:48Q&A
related talks
40mDeveloper Productivity Engineering for Open-Source Communities
45mAn Open Source Approach to Driving Developer Efficiency
34mData pipelines powered by Open source and fun
38mAutomotive Grade Linux - Evolution and Lessons Learned From 10 Years of Community Management
07mGitOps based CI/CD
27mAGL Update
43mDriving OSS Contribution: AGL OSPO-EG Insights and Strategic Plans...
37mAutomotive Grade Linux (AGL)