Unable to create EC2 Resource

Question

I am getting following error in my data pipeline

`We currently do not have sufficient t1.micro capacity in the Availability Zone you requested (ap-southeast-2a). Our system will be working on provisioning additional capacity. You can currently get t1.micro capacity by not specifying an Availability Zone in your request or choosing ap-southeast-2b. (Service: AmazonEC2; Status Code: 500; Error Code: InsufficientInstanceCapacity
`

This is how the pipeline definition looks for the EC2 resource
```
{
            "subnetId": "#{mySubnetId}",
            "securityGroupIds": "#{myEc2RdsSecurityGrps}",
            "instanceType": "#{myEC2InstanceType}",
            "name": "Ec2Instance",
            "actionOnTaskFailure": "terminate",
            "actionOnResourceFailure" : "retryAll",
            "maximumRetries": "3",
            "imageId": "ami-x",
            "id": "Ec2Instance",
            "type": "Ec2Resource",
            "terminateAfter": "2 Hours",
            "useOnDemandOnLastAttempt": "true"
        },
```

- I am required to specify subnetId - which ties it to a specific Availability Zone. 
- It only attempts creating resource once. Does not retry even after specifying `"maximumRetries": "3"`
- Even though `useOnDemandOnLastAttempt` is specified, it does not request on demand instance - guess because its not attempting more than once.

Vitor Castellani · Answer

Hello!

**AWS sometimes runs out of capacity in specific Availability Zones for certain instance types.**
Your pipeline's retry logic isn't effectively handling this issue, despite the "maximumRetries" configuration. Furthermore, although you've set "useOnDemandOnLastAttempt", it seems the system isn't attempting to use an on-demand instance, possibly because it's not retrying as expected.

A solution would be to either create/use a subnet in a different Availability Zone where there's available capacity or to investigate why the retry logic is not working as intended.

Abhishek Vasisht · Answer

Hi,

This error means that AWS lacks enough compute resources for spot and OnDemand instance  in the ap-southeast-2a availability zone. Wait for some time and try again, or you can create an instance in another AZ.   If you are running mission critical applications to avoid such scenarios you need to purchase dedicated instance https://aws.amazon.com/ec2/pricing/dedicated-instances/

If you have a multi-AZ Load Balancer and auto-scaling enabled solution design, it will launch new EC2 instances in different AZs for you.

Check out this solution for detailed information: https://repost.aws/knowledge-center/ec2-insufficient-capacity-errors

Thanks,   
Abhishek

Unable to create EC2 Resource

Relevant content