Zero-Regression Network CI/CD for Finance-Grade Reliability
summary
Claims elimination of post-deployment network issues by running a full virtual replica of production with the same playbooks, daily CI/CD runs, and failure injection; pipeline uses Jenkins, Vagrant/Libvirt, Ansible/Terraform, Robot Framework + Pabot, iperf and tc for latency and MTU checks; it accumulates 27,000 tests (growing), encodes past incidents as tests, and gates releases so only green commits reach production, enabling frequent zero-touch pushes and catching regressions before any trading impact.
AI-written summary · about
chapters (17)
- 0:00Zero-regression network CI/CD intro
- 1:41Finance-grade reliability matters
- 3:32Cost of manual network changes
- 5:18Zero regressions as goal
- 6:10Network config in Git
- 7:12Virtualized network lab
- 9:00Virtual environment pipeline
- 12:31Connectivity policy tests
- 15:01Failure scenarios testing
- 19:23Lab monitoring and alerting
- 20:27Test exceptions for weak spots
- 21:32Netm as CI/CD backbone
- 25:28Automated testing tools
- 26:34Safe deployment branching
- 28:03Fixing prod with automated tests
- 32:45Challenges and improvements
- 35:55Key takeaways & Q&A
related talks
27mContinuous Reliability. How?
35mCI/CD and devops with Ruby and Rails
26mIntelligent Deployment Pipelines
46mChaos Experiments - Datacenter Stress Testing
16mTesting All The Way To Production
01h 12mPitfalls of Continuous Deployment
28mUnderstanding CI/CD in Networking Context
25mContinuous Delivery and Testing in Production