PrepZone Logo
PrepZone

From Commit to Customer: The Delivery Lifecycle

Plan, code, build, test, release, deploy, operate, monitor — the loop every team runs.

Why this matters

  • Mapping the lifecycle exposes where your team waits on humans instead of automation.
  • The BookStore deployment journey in later modules follows these exact stages.
  • Production incidents often trace back to a skipped test stage or manual deploy step.
  • Senior interviews expect you to name stages and artefacts at each hop.
Git repoSource of truth
Maven buildJAR artefact
Docker imageImmutable package
CI/CD pipelineTest + push
KubernetesScale + route
ObservabilityMonitor + alert
The same Spring Boot app progresses through every module — from source code to production operations.

Eight stages, one continuous loop

When BookStore adds full-text book search, work starts in planning, lands in Git, becomes a Maven artefact and Docker image, passes automated tests, receives a semver tag, rolls to staging, promotes to production, and is watched via metrics and logs.

Key ideas

  • Plan — scope, SLO impact, rollback plan
  • Build — mvn package produces a reproducible JAR
  • Release — immutable tag ties code to artefact
  • Monitor — alerts on search latency and error ratio

Where BookStore invests automation

Manual QA gates stalled releases until integration tests and contract tests ran in CI. Deploy steps are now pipeline jobs; only production promotion requires approval plus smoke tests against /actuator/health.

Key ideas

  • Test — unit, integration, container smoke in pipeline
  • Deploy — same image from staging to prod
  • Operate — runbooks for database failover
  • Feedback — Grafana dashboards inform the next sprint
Java
git checkout -b feature/book-search
git commit -m "feat: add book search API"
git push origin feature/book-search
# CI: mvn verify → docker build → deploy staging → smoke test

Quick recall

Everything you need if you only revisit this box.

  • Eight stages from plan through monitor form a closed loop.
  • Artefacts: source → JAR → image → running pods.
  • BookStore uses the same container image in staging and prod.
  • Monitor feeds back into plan via incident and metric reviews.
  • Skipped test stages are the most common root cause of bad deploys.
  • Automation targets build, test, and deploy first.

Test yourself

Answer these before moving on — recall is what makes it stick.