Busting Silos & Red Tape: DevOps in Federal Government
summary
A Carnegie Mellon SEI team turns a prototype-turned-production mess into repeatable delivery: CI/CD runs on an internal network while production parity relied on emailed configs, causing late defects and budget pain. Fixes include VPN-based continuous delivery into a customer staging environment, markdown checked into git and auto-built into HTML/PDF (documentation-as-code), a RESTful security controller that gates deployments and creates JIRA tickets, daily root-cause PdCA called PC-SAM, canary releases, and experiments with static analysis and NLP/ML on commits to cut security review time and reduce defects.
AI-written summary · about
chapters (10)
- 0:00Start
- 3:29Ready to go!
- 5:08We got stuck
- 5:55Conflict
- 6:38We Took A Step Back to
- 9:28Security Bottleneck
- 13:46Experimentation and Learning
- 19:54Empathy
- 24:17Top Five Takeaways
- 25:36Here's what I'm looking for help with...
related talks
48mWhen DevOps Meets Regulation: Integrating 'Continuous' with 'Government'
32mContinuous Delivery Sounds Great But It Won’t Work Here
34mYou only have to change one thing to do the DevOps
29mOn the Path to DevOps and Lean Government
22mDevOps between modernity and tradition: an approach adapted to the public service
47mArchitecting for DevOps and Continuous Delivery
30mOur Continuous Delivery Journey
41mThe Introduction of DevOps in Large-Scale Organizations