Explore strategies for implementing progressive delivery and statistical observation in continuous delivery pipelines. Learn how to build control and observability into your CD process, enabling faster releases with reduced risk. Discover implementation best practices, including where to place progressive delivery controls, balancing centralization and team autonomy, and identifying key metrics for attributing system and user behavior changes to deployments. Gain insights from successful companies like Netflix, LinkedIn, and Booking.com on how they leverage these practices to consistently deliver results. Understand the importance of "guardrail" metrics, scaling considerations, and building trustworthy systems. Examine concepts such as feature flags, experimentation, and the "build-measure-learn" loop to validate ideas and improve decision-making in your development process.
Deliver Results, Not Just Releases - Control & Observability in CD
Overview
Syllabus
Introduction
Agenda
Copy the deck
Who am I
Why Im here
Definitions
Control Observability
Control of CD Pipeline
The payload
Control
Feature Flags
Experimentation
Code
Observability
Main Body
LinkedIn
Exellent
Observability Metrics
Guardrail Metrics
Unintended Consequences
Build for scale
Trustworthiness
Diversity
Bookingcom
Transparency
Scale
Deploying
Agility
Safety Net
Circuit Breaker
Auto abort
Validate ideas
Build measure learn loop
Validate new ideas
Lukas Premier
Facebook Gatekeeper
Complexity
Undo Button
Conclusion
Deployment vs Release
Where to Put Controls
Where People Miss It
Audit Trail
Two pieces of data
Taught by
GOTO Conferences
