CI and CD are nothing new. But learning more about how to do them better is where we should be focusing our energy – the goal is after all to enhance our ability to deliver value.
6. 6
Mean time
to recovery
Deployment
frequency
Lead time
to change
Change
failure rate
7. 7
Time
Seconds Minutes Hours Days Weeks
Dev working locally
Pipeline
Production
compilation unit tests
linting integration tests
business metrics
customer feedbackincidents
technical metricsperformance tests
security scanning
46. 46
Logging,
monitoring and
alerting owned
by dev team
Zero-downtime
deployments
Two (or three)
environments
Continuous
Deployment
Decoupled
pipelines
Build once,
use everywhere
The pipeline is
not special
Synthetic
monitoring
Always be testing
Automated,
repeatable,
isolated tests
Pairing
The pipeline is
the only way
Roll forward
Gradual
deployments
Config is code
You build it,
you run it
Automated
deployments
Paved roads and
building blocks
Trunk-based
development
CFRs in pipeline
Deployment
≠ release
Pipeline is code
Automated
compliance
Verified
deployments