Team Guide to Software Releasability

Team Guides for Software

Foreword

Introduction

1. What is software releasability?
2. What constitutes a delivery system?
3. What does resilient delivery feel like?
4. Warning signs of software delivery debt
5. Why invest in software releasability?
6. Relationship to Continuous Delivery
7. What this book is (not) about
8. How to use this book
9. Feedback and suggestions

1.Treat your pipeline as a product for resiliency and fast feedback loops

1. Key Points
2. 1.1Make your pipeline the single route to production
3. 1.2Your pipeline is now a product: invest in it
4. 1.3Avoid simply retro-fitting CD into a CI server
5. 1.4Measure delivery to visualize flow and identify bottlenecks
6. 1.5Design the delivery system to evolve with your needs
7. 1.6Apply monitoring and logging to minimize issues and downtime
8. 1.7Scale the infrastructure to avoid pipelines queuing up
9. 1.8Scale the practices and pipelines to support growing usage
10. 1.9Care for pipeline testability and usability to encourage adoption
11. 1.10Build security into and around the pipeline
12. 1.11Get started!
13. 1.12Summary

2.Regularly restore your delivery system to build resiliency

1. Key Points
2. 2.1Investing in minimal recovery time pays off in engineering time
3. 2.2Store everything in version control
4. 2.3Store your pipeline definitions in source control
5. 2.4Treat your pipeline as code or configuration?
6. 2.5Repeatable Environment Provisioning
7. 2.6Rollback without (too much) pain
8. 2.7Include the delivery system in disaster recovery scenarios
9. 2.8Get started!
10. 2.9Summary

3.Ensure delivery system is operable to minimize downtime

1. 3.1Visibility and tracing
2. Semantic Versioning

4.Ensure both practices and infrastructure can scale to meet usage growth

5.Care for pipeline testability and usability to encourage adoption

6.Measure delivery to visualize flow and identify bottlenecks

1. 6.1Plan and monitor build capacity to avoid queues
2. 6.2Build latency/throughput (?) / responsiveness
3. 6.3Monitor Code Repositories / Forensics
4. 6.4Monitor Your Builds and Environments
5. 6.5Aligning Agents with Delivery Tasks vs Generic Build Machines (it depends on cost of agent provisioning + running)
6. 6.6TODO: reference Steve Smith’s talk/book for further info

7.Treat your pipeline as a value stream to tackle largest bottlenecks first

1. Key Points
2. 7.1Static (or Nearly) Value Stream Mapping
3. 7.2Dynamic Value Stream Mapping (in the Pipeline)
4. 7.3Pipeline stages, not environments
5. 7.4Single Pipeline, Single Binary
6. 7.5visualize (actual) workflow, pipeline = value stream map, identify non-technical bottlenecks, support global optimization instead of local optimization
7. 7.6Goal: Short and Wide pipelines, not long and narrow
8. 7.7Resources
9. 7.8Inclusive Pipelines
10. 7.9Choose your tools wisely - the hidden costs of “DevOps solutions” (e.g. Visual Studio)

8.Organize teams to promote build and release ownership

1. 8.1CDaaS
2. 8.2Clarify Build Team vs Product Team Responsibility (Build, Infra, Deploy)
3. 8.3Build Team
4. 8.4Knowledge Sharing in Product Teams
5. 8.5DVTs for Collaboration Between Product and Ops/Infra Teams

9.Appendix A: build security into and around the pipeline

Terminology

References and further reading

1. Introduction
2. Chapter 1 - Treat Your Pipeline as a Product

About the authors

1. Chris O’Dell
2. Manuel Pais

Conflux Books