production

The live environment where real users use the service, as opposed to development or testing.

In more detail

Production gets the strictest settings, the most monitoring and the most careful changes, because failures affect real users. Other environments, such as development and staging, exist so problems can be found before they reach it.

Examples

An Ansible project usually keeps production as an inventory group with its own group_vars and its own vault ID.
Production Kafka needs at least three brokers, dedicated disks and acks=all rather than one dev-container broker.
Production JavaScript bundles omit development-only warnings, so deprecation messages never reach users' consoles.
The live environment where real users are served. In DORA metrics, a change counts as delivered only when it is serving production traffic, and a failure counts only when it reached production.
Changes on a production host can interrupt real users, so they deserve more care: identify the distribution, preview package transactions and read the plan before applying it. An upgrade there can replace configuration or restart a running service.
next build then next start, or node server.js from the standalone output, runs Next.js in production mode.
In Terraform: Changes to production deserve saved plans, required approval and credentials nobody holds on a laptop. prevent_destroy and branch protection are common guards.
In this course: Production code should not know it is being tested: seams with sensible defaults let production use the real clock, generator or directory while tests pass in their own. A test must never reach production systems, real money or real customers.

Related terms

Taught in

References

Certification courses