The last step in the Continuous Delivery pipeline is deploying to production. This deployment is either triggered manually or, when sufficient verification and automated tests are implemented, automatically. The vast majority of companies use a manually triggered deployment. But even if the pipeline does not go all the way from the beginning, Continuous Delivery provides great benefits by automating all steps necessary.

The pipeline then only has two kick-off spots: the initial commit to the repository that triggers the execution, and the final deployment to production after all steps have been verified, manually and automatically.

In a container orchestration environment, deploying to production, that is, either deploying to a separate namespace or a separate cluster, happens in the same way as deploying to test environments. Since the infrastructure as code definitions are similar or ideally identical to the ones executed before, this technology lowers the risk of environment mismatches to production.