Productive development workflows require fast turnaround times as well as fast feedback. Automating repetitive tasks minimizes the times spent on build, tests and deployments. Zero-dependency Java EE applications supports fast feedback by minimizing build, publish, and deployment times.

It's important to define which category of errors will break the build. Developers should be aware that a build is either broken, due to legitimate errors, or passed, without anything to complain about. Warnings that have no effect on the build outcome have little value.

Data migration is another important topic to consider. Deploying stateless applications is comparably easy; what needs to be taken into account are the database schemas that need to match the application code. Rolling updates together with migration scripts, that rollout modifications in small changes, enable applications to be deployed with zero-downtime. Applications therefore need to support N-1 compatibility.

Continuous Delivery depends on a healthy team culture. It's not sufficient to implement just the technical necessities; all software engineers need to embrace the principles. Potential build issues, test results, software quality, and deployment statuses should be visible to the whole software team.

Continuous Delivery processes support to continuously improve the software. Verification steps that are added, such as automated software tests, run every time the application is built, enabling regression tests and avoiding specific bugs to happen twice. This of course requires developers to put effort into the quality improvement. The effort put into Continuous Delivery will pay off in the long run.

The following chapter stays in the field of software quality and will cover testing enterprise applications.