Continuous Delivery for Legacy Code
summary
A monolith of 5 million lines across Classic ASP, .NET, VBScript, VBA, JavaScript, T‑SQL and PL‑SQL arrives with no VCS, deployments by copy‑paste, and roughly 1 million missing DB lines (≈2,600 stored procedures). The team imposes version control, a CI/CD pipeline with artifact repository, adds gold‑master tests and reverse‑engineers the domain model, incrementally strangles the monolith and uses dark launching to compare live outputs; result: safe, bi‑weekly deployments and far fewer unexpected failures.
AI-written summary · about
chapters (27)
- 0:00Introduction
- 2:10Legacy Code
- 3:11Apocalypto
- 8:58Why not rebuild
- 11:21How did they get here
- 14:51Atomic Habits
- 18:06Colocated Team
- 19:09Principles
- 21:23Adaptive Adaptability
- 22:11Continuous Delivery
- 24:52Branching Strategy
- 26:08Version Control
- 27:01Product
- 27:41Code Carter
- 28:48Java
- 29:33Other tools
- 32:01Repost
- 33:15Gold Master Test
- 34:58Cheat Sheet
- 36:39Refactoring
- 37:43Conquer the fear
- 38:38Reverse Object Mother
- 39:59Organizing Legacy Code
- 42:50Strangle Application
- 44:27Load Balancing
- 45:36Remove obsolete code
- 46:05Add new analytics
related talks
49mContinuous Delivery for Legacy Code
54mCtrl-Alt-Del: learning to love legacy code
23mDevOps and Legacy Code
50mContinuously delivering software to big brands
34mEnabling a Continuous Deployment Mindset
32mContinuous Delivery Sounds Great But It Won’t Work Here
26mLegacy Codebases are a DevOps Issue
48mContinuous Delivery: Tales from Windowsland