Continuous Delivery with Docker and Jenkins

Sample architecture

We already know that there should be slaves, (possibly multiple) master(s), and that everything should be duplicated into the test and production environments. However, what would the complete picture look like?

Luckily, there are a lot of companies that published how they used Jenkins and what kind of architectures they created. It would be difficult to measure if more of them preferred vertical or horizontal scaling, but it ranged from having only one master instance to having one master for each team.

Let's look at the example of Netflix to have a complete picture of the Jenkins infrastructure (they shared it as the planned infrastructure at Jenkins User Conference San Francisco 2012):

They have test and production ...