Case study of a good enough architecture
summary
Argues a practical “good enough” architecture that prioritizes fast feedback and cost effectiveness by using continuous delivery (deploy artifacts in ~10 minutes vs monolith builds 40–60+ minutes), microservices for small, fast pipelines, and heavy automation (Jenkins Job DSL, Ansible, IaC). Recommends unit tests and critical-path checks because ~90% of pipeline failures come from tests, aims to fix ~90% of production incidents within 30 minutes, and uses consumer-driven contracts (Spring Cloud Contract), APM/monitoring (NewRelic/Dynatrace) and conventions over configuration plus feature switches to make rollbacks and safe-to-fail behavior practical.
AI-written summary · about
chapters (27)
- 0:00Introduction
- 2:00Is there a perfect architecture
- 6:00Fast feedback
- 6:35Agile vs Kanban
- 9:10Deployment vs Release
- 11:00Biggest challenge
- 12:35Acceptance tests
- 13:25Unit tests
- 16:20Microservices
- 18:40Communication
- 19:40Communications
- 21:05Quality
- 27:40Versioning
- 28:20Performance
- 33:31Metrics
- 35:01Datadriven
- 35:33Return on invested
- 37:18Safe to fail
- 40:09repeatability
- 40:51automated tests
- 41:27automated deployments
- 41:42automated infrastructure
- 43:04backup
- 44:14example
- 46:04IP address
- 46:28Summary
- 48:46Thank you
related talks
46mArchitecture & DevOps
33mSoftware development practices for companies that want to make an impact
55mArchitectural Patterns for Rapid, Reliable, Frequent and Sustainable Development
01h 00mAgile Architecture
57mBuilding Systems that are #neverdone
41mSoftware Architecture in a DevOps World
40mSmoothing the continuous delivery path- a tale of two architectures
40mContinuous Delivery Anti-Patterns