The Linear Deployment Experience
summary
Claims that deployments must stay linear as scale increases and shows why they break: a Django/Postgres atomic transaction split across shards collapses a ledger and triggers a cascade into CQRS/Saga workarounds. Cites tools and patterns: Temporal for durable workflows, Argo/Octopus for progressive delivery, and describes building a release-orchestration framework called Quantum framework that handles merge conflicts, dependency graphs, traffic draining and progressive rollouts. Reports 50 companies surveyed with only 3 deploy daily and many struggling to deploy more than once per week.
AI-written summary · about
chapters (8)
- 0:00Linear deployment intro
- 5:01Motivation for linear dev
- 5:44Power of abstractions
- 9:44Order in distributed ops
- 14:26When deployment breaks
- 19:10Deployment as distributed problem
- 21:15Release orchestration solution
- 23:02Closing and key takeaways
related talks
38mJust Ship It!
34mEnabling a Continuous Deployment Mindset
40mContinuous Delivery Anti-Patterns
26mThe emotional chaos of deploying code
56mBrownfields DevOps in Practice
33mSoftware development practices for companies that want to make an impact
32mHow Your Deployment Strategy Influences How People Code
56mScaling Up!