How to set CPU and Memory on Task level and Container level
Ognif
asked 2 years ago

Question

I got the information to set my memory according to task level and to container level. Iam on EC2TaskDefinition and I create my tasks with CDK in typescript.
How to set cpu and memory inside EC2TaskDefenition, I found cpu and memory only in addContainer not in Ec2TaskDefinition construct.

Jordan Sandri · Accepted Answer

Hello!

For setting CPU and memory, you can actually do that within the TaskDefinition construct ([see TaskDefinition Documentation](https://docs.aws.amazon.com/cdk/api/v2/docs/aws-cdk-lib.aws_ecs.TaskDefinition.html)) which is a more general class. Just specify the **cpu** and **memoryMiB** configurations directly there. Then, if you want to run the task on an EC2 cluster, set the compatibility field to **Compatibility.EC2** ([Compatibility Documentation](https://docs.aws.amazon.com/cdk/api/v2/docs/aws-cdk-lib.aws_ecs.Compatibility.html)).

Hope this helps! 😊

re:Post Agent · Answer

When using the EC2TaskDefinition construct in CDK with TypeScript, you can set CPU and memory at both the task level and container level. Here's how to do it:

1. Task Level:
   For EC2TaskDefinition, you can set the CPU and memory at the task level when creating the task definition:

```typescript
   const taskDefinition = new ecs.Ec2TaskDefinition(this, 'TaskDef', {
     cpu: '256',
     memoryMiB: '512',
   });
   ```

The `cpu` parameter is specified in CPU units or vCPUs, while `memoryMiB` is specified in MiB.

2. Container Level:
   You're correct that you can set CPU and memory when adding a container to the task definition:

```typescript
   taskDefinition.addContainer('MyContainer', {
     image: ecs.ContainerImage.fromRegistry('my-image'),
     cpu: 256,
     memoryLimitMiB: 512,
     memoryReservationMiB: 256,
   });
   ```

Here, `cpu` is in CPU units, `memoryLimitMiB` is the hard limit in MiB, and `memoryReservationMiB` is the soft limit.

It's important to note that for EC2 launch type, you must specify either a task-level memory value or a container-level memory value. If you specify a task-level memory value, the container-level memory value is optional.

When using EC2 launch type, you have more flexibility with CPU and memory values compared to Fargate. Any value can be used for CPU, and the memory can be set to any amount that fits within the resources of the underlying EC2 instance.

Remember that the total CPU and memory reserved for all containers within a task should not exceed the task-level CPU and memory values. If you're using task-level resource allocation, ensure that the values you set are appropriate for the EC2 instance types in your ECS cluster.  
**Sources**  
[class CfnTaskDefinition (construct) · AWS CDK](https://docs.aws.amazon.com/cdk/api/v2/docs/aws-cdk-lib.aws_ecs.CfnTaskDefinition.html)  
[class TaskDefinition (construct) · AWS CDK](https://docs.aws.amazon.com/cdk/api/v2/docs/aws-cdk-lib.aws_ecs.TaskDefinition.html)  
[Troubleshoot Amazon ECS task definition invalid CPU or memory errors - Amazon Elastic Container Service](https://docs.aws.amazon.com/AmazonECS/latest/developerguide/task-cpu-memory-error.html)

Ognif · Answer

There is definitly no cpu and memory in the Ex2TaskDefinition construct. 
https://docs.aws.amazon.com/cdk/api/v2/docs/aws-cdk-lib.aws_ecs.Ec2TaskDefinition.html

How to set CPU and Memory on Task level and Container level with EC2TaskDefintion and CDK

Add your answer

Relevant content