Today we launched #ECS Capacity Providers at #reInvent. Capacity providers change how you think about and use compute capacity to run containers in ECS. 1/n aws.amazon.com/about-aws/what…
With capacity providers, you can do the following new things in ECS: (1) create a service that uses a mix of on-demand and spot capacity, at a predefined ratio for the service. For example, you can run 2/3 of the tasks on EC2 On-Demand instances and 1/3 on EC2 Spot. 2/n
(2) Run a service that is automatically balanced across multiple availability zones, and the balance is strictly enforced by the service scheduler. 3/n
(3) Enable Cluster Auto Scaling, where ECS manages scale-in and scale-out of your ASG automatically, including scale to and from zero, and instances running tasks will not be selected for scale-in 4/n
(4) Use Fargate Spot and Fargate in the same service, again at a predefined ratio (or go all in on Fargate Spot). 5/n
(5) Create IAM policies and roles that allow or deny access to specific compute infrastructure within a cluster - enhances security, compliance, and cost control. 6/n
If you are at #reinvent, sign up for session CON325 “Enabling Application-First with Amazon ECS Capacity Providers” on Wednesday or Thursday. We have a lot of plans for capacity providers over the coming months and years, this just Day 1! 7/7
