Contributor Experience matters
summary
Apache Airflow describes concrete results: switching from Travis to GitHub Actions plus self-hosted runners, Docker image caching, GNU parallel and selective matrix runs cuts CI from over an hour to about 15–20 minutes and yields "up to 90% speedups." Security controls only let committers use self-hosted runners, workflows cancel duplicate jobs, and heavy caching uses GitHub Package Registry; upkeep costs run roughly $500–$600/month on AWS. Remaining pain points: runner maintenance, environment cleanup between jobs, and GitHub’s evolving approvals and ephemeral-runner support.
AI-written summary · about
chapters (9)
- 0:00Introduction to project
- 3:39Apache Airflow overview
- 8:51Types of contributors
- 13:42Contributor workflow challenges
- 19:46CI as key component
- 23:38Moving from Travis CI
- 26:08GitHub Actions optimization
- 35:40CI infrastructure costs
- 38:54Security and approvals
related talks
31mThe Silent Symphony: Keeping Airflow's CI/CD and Dev Tools in Tune
46mQuality Beyond Testing: Jenkins to GitHub Actions Migration: Optimized Processes & QA
30mSpend More Time Coding & Automate the Boring Stuff
55mGitHub Actions - CI:CD and more
31mGetting Started with CI/CD using Github Actions
50mGitHub Actions in action
36mGitHub Actions: Open Source Workflow Automation
37mHandling User Feedback: A Case Study With GitHub Actions Manager