In this post, we’re going to add a scaling policy to our WordPress auto scaling group.
I read somewhere that all AWS VMs should be launched in an auto scaling group, and at the time that seemed fairly eccentric to me. But after having done my Bastion host template, I agree with that statement wholeheartedly. I don’t recall ever saying I’d like to stand up a server, but if it dies, I want it to stay dead. Generally, you’re standing up a server so it will be available, listening for requests, and sending responses. You don’t know when the request might come, but whenever it comes you want this server to be ready for it.
By itself, an auto scaling group is just a wrapper for one or more VMs that makes them self healing. And who doesn’t like self healing hosts? That said, if you really want high availability, and you expect a load on the servers that is not entirely predictable (and how many servers have an entirely predictable load), then an auto scaling group by itself isn’t enough. You also need to apply a scaling policy to it, so it can provide on demand capacity, which is a central theme in cloud computing. That’s what we’re going to do now.