Now that we have all the basic services required for our WordPress installation to consume, we can make a start on deploying the compute resource to install WordPress on. This is where things get interesting, as we have to build logic into our playbook so that if our site is up and running, we can deploy updates to the operating system and roll out new images without any downtime.

But if it is a new deployment, we need to launch an instance, attach it to the Elastic Load Balancer, install the software stack, configure WordPress, and create an image we can then use in a launch configuration, which we will need to attach to an autoscaling group.

While this may seem complicated, building this logic into the playbook will make it a lot easier to maintain and hand over to someone else to manage/run, as they will not need to worry about the existing deployment, they will just need to run the playbook.