Continuous deployment automatically promotes infrastructure changes from a testing environment to production after required tests pass, removing the manual approval stage. section 7.1.3
Apply it when the delivery pipeline can reliably:
- Run thorough integration and end-to-end tests that provide confidence in system behavior. section 7.1.3
- Deploy changes directly to production without waiting for manual approval. section 7.1.3
- Recover failures quickly through a known rollback procedure or runbook. section 7.1.3 section 7.1.4
- Classify changes so that standard or emergency changes follow automation while major and new changes retain manual approval. section 7.1.4
Routine infrastructure change
A weekly server-tag update can use continuous deployment when its behavior is well understood, its tests are automated, and a runbook provides a safe rollback. Similar standard changes include upgrading orchestrator container images, deploying a queue, or adding a monitoring alert. section 7.1.4
Table 7.1 Types of changes and delivery approach (view table figure)
| Rolling back an operating system image to a previous version |
|||
Compare standard, emergency, major, and new infrastructure changes with their recommended delivery approaches and approval requirements. section 7.1.4
