Continuous Delivery - The Missing Parts
by Nicholas Tuck
1. "Until your pretty code is in production, making money, or doing whatever it does, you've just wasted your time"
1.1. - Chris Read @cread #LondonCI
2. Dark Launching
2.1. test systems in small gradual way to get users to use it before you do a full blown launch of that feature
2.2. * Look into more
3. Understanding Work
3.1. Business Projects
3.1.1. search, UI improvements
3.2. Internal Projects
3.2.1. architecture changes
3.3. Changes
3.3.1. deployments, schema updates
3.4. Unplanned Work
3.4.1. downtime, investigations
4. Need Feedback Loop from Ops team back to Development
4.1. Must be constructive feedback
4.2. Culture of Continual Experimentation and Learning
4.2.1. Multivariant testing
4.2.2. use operations data to drive development decisions
5. Can Adopt DevOps through
5.1. Culture
5.1.1. People and Process are first
5.1.2. not about management dictating
5.2. Automation
5.2.1. Manual systems and processes will always stagnate, wither and die
5.3. Measturement
5.3.1. Response times
5.3.2. Usage stats
5.4. Sharing
5.4.1. Culture of feedback
6. Value Stream Map
6.1. cycle time of something to production
6.2. improving "Flow"
7. Immutable Infrastructure
7.1. Don't need access to production systems
7.2. Need logs? Lets get a log viewing solution in place
7.3. Or new term
7.3.1. Disposable Infrastructure
8. Orchestration
8.1. Run Deck . org
9. Speaker
9.1. Paul Stack
9.2. @stack72
9.3. [email protected]
9.4. Infrastructure Engineer for a cool startup
10. Definition
10.1. set of practices and principles aimed at building testing and releasing software faster and more frequently
10.2. 8 principles
10.3. 4 practices
10.3.1. Build binaries only once
10.3.2. Use precisely the same mechanism to deploy to every environment
10.3.3. Smoke test your deployment
10.3.4. If anything fails, stop the line!