About autoscaling, I have recorded video and written a blog post a few months back. At that time autoscaling was based only on thresholds, commonly called metric-based autoscaling. Precisely speaking, it means that after crossing some defined level of CPU or memory utilization, you could expect the automatic provisioning of additional compute instances defined in your instance pool (scale-out process). When the CPU load was smaller or memory has been freed, a number of the machines should drop down to the expected level (scale-in). Recently compute autoscaling has been enriched with a schedule-based scenario. It means autoscaling events take place at the specific times that you schedule. It could be a very busy “end of the year” where your standard computing power is just not enough and you need to prepare some bigger scale setup. Also, it can be any other circumstances that can be predicted and you can define time-based scale-out and scale-in policies for the autoscaling capability of your instance pool. Below you can find a fresh video that showcases the usage of Terraform for schedule-based autoscaling, deployed in my OCI tenancy. Hope you will find this dish tastefully. You can of course cook by yourself with this GitHub repo. 🙂
Martin, The Cook.