How We Introduced Continuous Delivery?
Download as PPTX, PDF0 likes427 views
The document discusses how a team introduced continuous delivery for a legacy application and a new application. For the legacy application, the cycle time was reduced from over 35 days to around 10 days. For the new application, they achieved 1-5 releases per week with over 40 releases to production in 5 months and a minimum cycle time of 2 hours. The key aspects that enabled this included forming a dedicated team, focusing on the value stream, implementing kanban/lean principles, automation including full end-to-end pipelines, and eating their own dog food by applying continuous delivery principles to their own improvement process.
How We Introduced Continuous Delivery?
- 1. How we introduced Continuous Delivery? Alex Schwartz www.flickr.com/photos/burtonandy/4702810536/ Places Backend Delivery Team
- 2. How we introduced Continuous Delivery? Old Days: • Cycle time >35 days • 4-6 releases per year Now – for a legacy app: Cycle time ca. 10 days New App: PBAPI • 1-5 releases per week • 40 releases in prod in 5 month • Min cycle time 2 hours • Average cycle time ½ day www.flickr.com/photos/mc_masterchef/178245348/
- 3. Focus – Form a Team Do not underestimate Conway’s Law. Release Dev QA Mgmt OPS Build Mgmt
- 5. Focus on the value stream Main Metric: Release Cycle Time
- 6. Eat your Own Dog Food You understand the value of Continuous Delivery? Use the same principle for your Improvement Process. Baby Steps, Baby Steps, Baby Steps , Baby Steps…. www.flickr.com/photos/mshades/2132689629/
- 7. Automation “ Automating the wrong thing to early is dangerous Bullshit ! • Automate everything you do twice. • Automation will be your rescue. www.flickr.com/photos/gibbisons/5538845069/
- 8. Automation Click-Button Deployments Our Automation: Full E2E pipeline in one tool, from commit to prod (for Thor systems).
- 9. Further Topics … TDD DevOps QA Test Automation Cut your ATD Acceptance Criteria stories Pairing Trust People