As promised I will be going more into ECS this post. First off the ECS cluster, this utilizing the two methods that I mentioned before EC2 or Fargate, Fargate is typically the best option as it is serverless and is much easier to scale as you do not need to manage the underlying instances. The ECS cluster you create can run both tasks and services, services are long-term processes that need to be running consistently, such as a web server. Tasks are used for short-term processes such as a batch job. When it comes to scaling your ECS cluster, the tasks are treated separately from the EC2 instances those are scaled separately from each other, this is not an issue if you are using Fargate as that is a serverless offering. There are 3 different way you can scale in ECS. First off we have target tracking, this looks for a specific CloudWatch definition and scales up when that definition is met. Next up we have step scaling, which will increment up as certain CloudWatch definitions are met. Finally we have scheduled scaling which is a good option if there is a specific time or date usage goes up exponentially and you can better prepare. You can either scale with an ASG, or you can utilize the ECS cluster Capacity Provider which attached to an ASG and scales your tasks and instances automatically based on usage.