Why your best engineers should look after the worst systems
summary
Claim: systems with 70 monoliths and only single-digit test coverage require senior engineers because changes need archaeology, readable tests and operability fixes. The project lifts and shifts a zip-file repo into the cloud using a traffic slider and testing in production while adding observability, feature flags and small iterative Strangler rewrites. Expect months to add a simple field, rely on experienced pairs, form small autonomous teams, prioritize operability work alongside bugs, and measure small wins (build reliability, deploy time, failure rate) to justify continued investment.
AI-written summary · about
chapters (10)
- 0:00Legacy systems and their value
- 2:24Defining legacy code
- 11:10Maintenance death spiral
- 15:29Language shapes legacy view
- 19:27Experience for legacy maintenance
- 21:10Autonomous teams for legacy
- 25:14Rewriting legacy is bad
- 27:45What legacy teaches you
- 30:43Measuring legacy success
- 32:10Q&A: Testing legacy code
related talks
50mTactics and strategies on software development: how to reach ...
43mEffective CI/CD for Large Systems
41mThe monolith must die! - How to convince managers
23mDevOps and Legacy Code
01h 03mBeyond the innovator’s dilemma
55mGetting Unstuck Strategies for Surviving in a Rapidly Changing Post-Agile World
20mDoing Things the Hard Way
01h 27mRedesigning Legacy Systems : Keys to Success