Changing Engines Inflight - A Real Life Modernisation Story
summary
A 28‑year stockbroking platform modernizes without downtime by running two parallel tracks: stabilize legacy with Grafana/Prometheus monitoring and centralized logging, then build a new backend in Go and Python with reusable recipes and startup projects. Migration uses Strangler fig refactors, API split for mobile, DNS-based cutover, and shadow traffic and diffing for nightly batch and transactional validation until outputs match; force‑upgrades, rollout controllers and rollback playbooks handle failures. Migration succeeds but surfaces missed edge cases and learning about documentation and incident ownership.
AI-written summary · about
chapters (7)
- 0:00The legacy challenge
- 2:52Modernization constraints
- 6:04Stabilizing existing systems
- 9:07Planning the new architecture
- 13:56Migration design patterns
- 20:46Transactional migration
- 27:58Learnings and key takeaways
related talks
29mHow to Replace a Jet Engine of your System In Flight
20mLegacy Migration... Differently
30mApplication Modernization Using Cloud Strangler Pattern and DevOps Best Practices
58mLegacy modernization made practical: A live coding example of modernization patterns
53mHow to migrate a mission critical railway system
35mPerforming Infrastructure Migrations at Scale
40mModernizing the Legacy
43mBeyond Strangler Fig: Effective Patterns for Architectural Modernization