My notes from the DevOps Handbook

by Gene Kim, Jez Humble, Patrick Debois, John Willis

Focus on deployment lead time

• subset of the value stream
• engineer - anyone working in the value stream
• start: check in of a change into VCS
• end: successfully running in production
• producing value to the customer
• feedback and telemetry
• first phase of work - design and development
• uncertain, variable processing time
• second phase of work - testing and ops
• goal to minimize variability - short and predictable lead time

Our goal is to have testing and operations happening simultaneously with design/development

• small batches
• less sequentiality
• quality in every part of the value stream
• for example TDD

Lead time vs Processing Time

• Lead time starts when the request is made
• Processing time starts when we begin to work
• Focus on the Lead Time -> experience of the customer
• Process Time / Lead Time ratio

Scenario: Deployment Lead Time Requiring Months

• large, complex organizations
• tightly coupled, monolithic applications
• reliance on manual testing
• heroics at every stage of the Value Stream
• days of fixing and investigation

DevOps Ideal: Deployment Lead Time of Minutes

• fast and constant feedback to developers
• quick and independent implementation, integration and validation
• deployment to production
• small code changes in VCS
• automatic tests
• modular architecture (see Uncle Bob's Clean Architecture)
• encapsulation
• loosely-coupled

%C/A as measure of rework

• percent of Complete and Accurate
• quality of the output of each step in the value stream
• percentage of work that is 'usable as is' by customers