DevOps Key Practices
What this domain covers
DevOps Key Practices tests the technical practices that make flow and feedback real: continuous integration, continuous delivery and deployment, infrastructure as code, configuration management, automated testing, version control and monitoring. Questions often distinguish similar practices.
Continuous integration (CI)
Developers merge small changes to a shared trunk frequently, and every merge triggers an automated build and test run. The purpose is early detection of integration defects. CI is not a nightly build and not simply a build server.
Continuous delivery vs continuous deployment
Continuous delivery keeps every change in a releasable state; releasing to production remains a business decision. Continuous deployment automatically releases every change that passes the pipeline. The exam frequently tests this distinction.
Infrastructure as code and configuration management
Environments are defined in versioned, declarative files and recreated on demand. Configuration management keeps systems in a desired state and is idempotent. Immutable infrastructure replaces servers rather than patching them, reducing drift between staging and production.
Testing, version control and monitoring
Use a test pyramid: many fast unit tests, fewer service tests, very few UI tests. Shift testing left, then verify in production. Put code, configuration, pipeline definitions and documentation under version control. Monitor user-facing outcomes, not only server CPU.
Worked example
A team spends two days manually preparing each release environment, and staging differs from production. They codify the environment, version the definitions and rebuild staging from the same source. Setup drops to 20 minutes and a large class of works in staging defects disappears. Reasoning: the environment was an undocumented dependency; making it code removed variance and enabled continuous delivery.
Traps that cost marks
- Calling any automated build CI when the build is allowed to stay red.
- Claiming continuous delivery while releases still require a manual multi-day handoff.
- Automating a process that should be removed first.
- Monitoring infrastructure only, with no signal on user experience or error rates.