Amazon ECS Cluster Auto Scaling with a Capacity Provider

You can now scale your Amazon Elastic Container Service (ECS) clusters (as well as your services) automatically. In this AWS video tutorial, I show you how the new Capacity Provider feature enables Amazon ECS clusters to work with Amazon EC2 Auto Scaling to scale the ECS instances automatically based on load.
For more details on Cluster Auto Scaling, check out the FREE Cheat Sheets from Digital Cloud Training: digitalcloud.training/certifi...
This video lesson is an excerpt from the comprehensive training course for the AWS Certified Developer Associate from Digital Cloud Training. To access the full course, visit: digitalcloud.training/aws-cer...
If you find this helpful, please SUBSCRIBE to our channel.
At Digital Cloud Training, it's our mission to help you achieve success in your cloud career. To learn more about our popular on-demand courses for Amazon Web Services, visit digitalcloud.training/aws-tra...
Apply coupon code "youtube" for a 10% discount.
If you have any questions feel free to leave a comment!

[@hopewise99]

Where is the last lesson about scaling tasks using ECS autoscaling service?

[@StephenRayner]

Would love to see this done in Terraform

[@sb8508_1]

How do you register those new EC2s to the Target groups?

[@atomiorootshell2]

Thank you 😉

[@akshayvit1312]

Thank you very much for such an informational video. it was very helpful. I understood when the desired count is increased ECS containers also increased. However when we reduce desired count. ECS container instances remain the same and capacity unutilized. Could guide me to achieve scale down ECS instances when capacity not utilized.

[@DigitalCloudTraining]

Use Auto Scaling: docs.aws.amazon.com/AmazonECS/latest/developerguide/service-auto-scaling.html

[@raghum8767]

Is it possible to configure two ec2 different types of instances while auto scaling in ECS? For example existing ecs is c5.large, while auto scaling it should run a lower instance

[@DigitalCloudTraining]

Not exactly, you would need separate ASGs launch templates.

[@SamualAnthony]

Hi, I have purchased a subscription to your course. But I cannot find this lesson, or the previous lesson you have mentioned in this video

[@DigitalCloudTraining]

You can get the full video course here: digitalcloud.training/aws-certified-developer-associate/

[@adamdaniel5766]

Hi where can I foud the last section that you are talking about

[@DigitalCloudTraining]

You can get the full course here: digitalcloud.training/aws-certified-developer-associate/

[@sambasiva1702]

can you please link the last lesson?

[@DigitalCloudTraining]

You can view the entire playlist here: th-cam.com/video/r4wBvN4Vebc/w-d-xo.html
They are excerpts from this full course: digitalcloud.training/aws-certified-developer-associate/

[@pradeep_kumar_ind]

Thanks for this video. Very useful. I have one query. Why did you mention the desired capacity as 4? It should be 6 if you want to run 8 tasks, because, the first AG took care of adding 2 instances already and the capacity provider should add 6 instances to making it total 8!

[@DigitalCloudTraining]

This is cluster auto scaling not service auto scaling. It's about adding the underlying cluster instances not the number of tasks.

[@pradeep_kumar_ind]

@@DigitalCloudTraining I do understand the difference, as I myself work on cloud. What I wanted to understand is, in this example there is one t2.micro EC2 instance per task. That means if you are creating 8 tasks, the cluster should have 8 EC2 instances. 2 instances have already been added through EC2 autoscaling, but you are adding only 4 EC2 instances through ASG associated with that Capacity Provider. So, I am still confused, how would scaling-out tasks will create 8 instances?. Thanks.

[@DigitalCloudTraining]

@@pradeep_kumar_ind the number of instances is the number of underlying hosts running the tasks, the tasks are individual Docker containers. You can have more tasks than you have underlying instances as you can run many tasks on each cluster host.

[@pradeep_kumar_ind]

@@DigitalCloudTraining In your video, between 2:50 and 2:58, listen to what you have mentioned. You said one task per instance. I asked you this query based on what you said. :)

[@shiv4667]

@@pradeep_kumar_ind and he never replied