All field notes

Production7 min read

A green build is not the same as production readiness

Compilation proves that source can become an artifact. Production readiness also requires operability, security, rollback, observability, and evidence that real journeys still work.

01

Define the release boundary

Before deployment, identify the exact artifact, configuration, migration set, feature flags, and external services that form the release. A passing build for a different commit is not evidence for the artifact being promoted.

The release record should connect approved requirements and QA evidence to that immutable build.

02

Prepare for failure before launch

A deployment plan needs health checks, logs, metrics, ownership, rollback or roll-forward steps, and thresholds that determine when action is required. These are product safeguards, not infrastructure trivia.

  • Verify secrets and configuration without exposing them.
  • Exercise migrations and recovery paths.
  • Define health and business-success signals.
  • Assign an owner for the launch window.
03

Production is another evidence source

Synthetic checks can confirm availability, while real telemetry shows whether people complete important journeys. Compare both against expected behavior and privacy constraints.

A release is complete when the new version is healthy, observable, and understood—not merely when the deployment command exits successfully.