Automation and safe delivery technology
Automate to remove friction, not for its own sake
Automation is one pillar of DevOps, not the whole of it. Its purpose is to make the right thing the easy thing: repeatable, fast and low-risk. Automating a broken process simply produces failures faster, so simplify first, then automate.
Technology areas a candidate must know
- Version control as the single source of truth for code, configuration and documentation; reviewable source definitions should be traceable. Secrets, sensitive datasets and generated outputs require appropriate separate handling, not indiscriminate commits.
- CI/CD pipelines: automated build, unit and integration tests, security and quality gates, artefact repositories, automated deployment and release orchestration.
- Infrastructure as Code (IaC): machine-readable definitions or provisioning instructions, often declarative, versioned and reviewed like application code; enables repeatable, disposable environments and immutable infrastructure.
- Configuration management tools keep many servers consistent and idempotent.
- Containers and orchestration package applications with dependencies and manage scheduling and scaling.
- Cloud and elastic infrastructure allow self-service environments on demand.
- Architecture for flow: loosely coupled, independently deployable services with clear contracts reduce coordination cost.
- Automated testing: a balanced test portfolio (many fast unit tests, fewer integration, selective end-to-end and performance tests) gives fast, trustworthy feedback.
Traps in questions
- "Automate everything" as an unquestioned good: the real target is toil - repetitive, manual, non-value-adding work. Automating a one-off task can cost more than it saves.
- "Buying a pipeline tool means adopting DevOps": technology is an enabler, not the definition.
- Treating IaC as documentation rather than executable, tested code.
- Assuming automation replaces the need for human judgement in releases and incident response.
- Forgetting approval and compliance steps can also be automated as policy checks in the pipeline.
Rule of thumb
If a manual step is frequent, error-prone and repeatable, automate it into the pipeline; if it is rare or judgment-dependent, evaluate its risk and economics before deciding what remains human and what can be assisted safely.