Continuous Deployment - The Last Frontier for Continuous Delivery
summary
Argues that Continuous Deployment is about running experiments in production not just automating deploys and shows concrete strategies with pros and cons: recreate/in-place swaps one version for another, simple but forces downtime; rolling updates replace instances in batches (example updates 2 of 10 instances at a time) to reduce impact; blue‑green duplicates environments for instant rollback at the cost of double resources; canary, shadow (traffic mirroring with Flagger), and circles route small percentages or multiple variants to real users to detect bugs early and stop bad variants, so require observability, backwards compatibility and a tested rollback plane.
AI-written summary · about
chapters (6)
- 0:00Start
- 2:32Continuous Integration
- 5:34The Hierarchy of Continuous Integration
- 9:55Pros and Cons for Your Rolling Deployment
- 11:35Pros and Cons
- 21:29Best Practices
related talks
31mFireside Chat: Continuous Deployment
41mAll Things Continuous
40mContinuous Delivery Anti-Patterns
26mThe Many Ways to Deploy Continuously
25mContinuous Delivery and Testing in Production
32mHow Your Deployment Strategy Influences How People Code
46mContinuous Deployment Workflows
24mWhy, What, and How of Continuous Delivery