Lessons Learned from Building a Deployment Process for 100+ Devs at a Large Bank
summary
Built a source-controlled schema workflow for 150+ developers across 12+ teams, replacing file shares and ad hoc scripts with Git, Azure DevOps, Octopus Deploy and Redgate; enforcement removes uncommitted test changes and requires feature-branch → PR flow. Result: 60% failure rate drops to near zero and two‑hour windows shrink toward 15 minute deployments. Lessons: define business rules first, iterate on the process, track deployment frequency/time/failure rate, and know when to say no to shortcuts that bypass the pipeline.
AI-written summary · about
chapters (10)
- 0:00Introduction
- 1:04The manual database deployment nightmare
- 5:36Lesson one automation doesn’t fix broken processes
- 6:42Lesson 2: Embrace change
- 9:56Lesson 3: Business rules first
- 13:03Lesson 4: Enforce consistency
- 14:28Lesson 5: Iterative evolution
- 15:11Lessons since rollout
- 18:59Lesson 6: Transparent ownership
- 20:04Q&A
related talks
47mContinuous Delivery Warstories
01h 01mA team's transition to Continuous Delivery
54mFrom Code to Customer
21mBuilding A DevOps Culture in a Remote World
52mFrom 0 $100 billion: Inside lessons on how to scale your dev workflow
51mPatterns & Playbooks to Make Every Commit Production Ready
49mContinuous Delivery for Legacy Code
41mContinuous delivery in 10 easy mistakes