Community Note

Voting for Prioritization

• Please vote on this issue by adding a 👍 reaction to the original post to help the community and maintainers prioritize this request.
• Please see our prioritization guide for information on how we prioritize.
• Please do not leave "+1" or other comments that do not add relevant new information or questions, they generate extra noise for issue followers and do not help prioritize the request.

Volunteering to Work on This Issue

• If you are interested in working on this issue, please leave a comment.
• If this would be your first contribution, please review the contribution guide.

Ive got the same issue. It seems that the default values are being set to null, but the default api values for desired_capacity_types for both CreateAutoscalingGroup and UpdateAutoscalingGroup is units

Update: any console changes to asg are setting desired_capacity_types value, so terraform detects this as a drift and tries to revert that to null, causing the error above

Ref on default values: https://docs.aws.amazon.com/autoscaling/ec2/APIReference/API_CreateAutoScalingGroup.html
https://docs.aws.amazon.com/autoscaling/ec2/APIReference/API_UpdateAutoScalingGroup.html

Hey @hans-d 👋 Thank you for taking the time to raise this! Can you supply a sample configuration that can be used to reproduce this and/or debug logging? That information will be useful for whoever picks this up.

This bug seems to have been here since instance requirement support was added. I have had problems with this for ages..

Also happens with AWS provider v5.51.1

can second @rctunisi here, ran into this when I modified the ASG in the console and tried to sync it in Terraform code.
AWS Provider version 5.47.0.

The only "fix" that can be done for non-critical-absolutely-testonly-workloads I can suggest is to terminate the entire ASG and recreate it and maintain it only in TF, while avoiding editing the capacity units or the number of nodes in the web console.

I received that error and got some debug logs to help in that case, seems that because desired_capacity is in ignore_changes for resource aws_autoscaling_group the value sent to AWS Autoscaling Group API is empty (&DesiredCapacityType=&MaxSize=3), stopping at API validation.

Scenario:

• Using directly Self Managed Node Group Module
  in a EKS not created by terraform cluster.

• I applied the terraform for node group, in the second moment updated the desired state by GUI then after that I can't update more using terraform.

Terraform Error:

╷
│ Error: updating Auto Scaling Group (*******************************): operation error Auto Scaling: UpdateAutoScalingGroup, https response error StatusCode: 400, RequestID: *******************************, api error ValidationError: 1 validation error detected: Value '' at 'desiredCapacityType' failed to satisfy constraint: Member must have length greater than or equal to 1
│ 
│   with module.self_managed_node_group.aws_autoscaling_group.this[0],
│   on .terraform/modules/self_managed_node_group/modules/self-managed-node-group/main.tf line 491, in resource "aws_autoscaling_group" "this":
│  491: resource "aws_autoscaling_group" "this" {
│ 

Debug Logs:

2024-06-12T00:15:34.634-0300 [DEBUG] module.self_managed_node_group.aws_autoscaling_group.this[0]: applying the planned Update change

2024-06-12T00:15:34.650-0300 [DEBUG] provider.terraform-provider-aws_v5.53.0_x5: HTTP Request Sent: http.request.header.amz_sdk_request="attempt=1; max=25" http.request.header.x_amz_date=20240612T031534Z net.peer.name=autoscaling.eu-west-2.amazonaws.com rpc.system=aws-api http.request.body="Action=UpdateAutoScalingGroup&AutoScalingGroupName=eks-*******************************&DesiredCapacityType=&MaxSize=3&NewInstancesProtectedFromScaleIn=false&Version=2011-01-01" http.request.header.content_type=application/x-www-form-urlencoded aws.region=eu-west-2 http.request_content_length=184 http.url=https://autoscaling.eu-west-2.amazonaws.com/ tf_aws.sdk=aws-sdk-go-v2 tf_rpc=ApplyResourceChange http.request.header.x_amz_security_token=***** http.request.header.amz_sdk_invocation_id=******************************* tf_aws.signing_region= tf_provider_addr=registry.terraform.io/hashicorp/aws http.user_agent="APN/1.0 HashiCorp/1.0 Terraform/1.5.6 (+https://www.terraform.io) terraform-provider-aws/5.53.0 (+https://registry.terraform.io/providers/hashicorp/aws) aws-sdk-go-v2/1.27.1 os/macos lang/go#1.22.2 md/GOOS#darwin md/GOARCH#arm64 api/autoscaling#1.40.10" tf_resource_type=aws_autoscaling_group http.method=POST rpc.service="Auto Scaling" tf_mux_provider=*schema.GRPCProviderServer tf_req_id=******************************* rpc.method=UpdateAutoScalingGroup @module=aws http.request.header.authorization="AWS4-HMAC-SHA256 Credential=********************/20240612/eu-west-2/autoscaling/aws4_request, SignedHeaders=amz-sdk-invocation-id;amz-sdk-request;content-length;content-type;host;x-amz-date;x-amz-security-token, Signature=*****" @caller=github.com/hashicorp/aws-sdk-go-base/v2@v2.0.0-beta.53/logging/tf_logger.go:45 timestamp=2024-06-12T00:15:34.650-0300

2024-06-12T00:15:35.756-0300 [DEBUG] provider.terraform-provider-aws_v5.53.0_x5: HTTP Response Received: rpc.service="Auto Scaling" tf_provider_addr=registry.terraform.io/hashicorp/aws http.response.header.content_type=text/xml http.response_content_length=397 tf_aws.signing_region= tf_req_id=******************************* @module=aws aws.region=eu-west-2 http.duration=1105 http.response.body="<ErrorResponse xmlns="http://autoscaling.amazonaws.com/doc/2011-01-01/"> <Error> <Type>Sender</Type> <Code>ValidationError</Code> <Message>1 validation error detected: Value '' at 'desiredCapacityType' failed to satisfy constraint: Member must have length greater than or equal to 1</Message> </Error> <RequestId>*******************************</RequestId> </ErrorResponse>" http.response.header.date="Wed, 12 Jun 2024 03:15:35 GMT" rpc.method=UpdateAutoScalingGroup @caller=github.com/hashicorp/aws-sdk-go-base/v2@v2.0.0-beta.53/logging/tf_logger.go:45 tf_rpc=ApplyResourceChange tf_aws.sdk=aws-sdk-go-v2 tf_resource_type=aws_autoscaling_group rpc.system=aws-api tf_mux_provider=*schema.GRPCProviderServer http.response.header.x_amzn_requestid=******************************* http.status_code=400 timestamp=2024-06-12T00:15:35.756-0300

Hello! My team experienced this, and we resolved it by adding the "desired_capacity" and "desired_capacity_type" properties to the Terraform "aws_autoscaling_group" resource, then modifying the resource for the lifecycle tags: ignore_changes = [desired_capacity, desired_capacity_type]. Basically adding the properties, which would be used upon resource creation, then ignoring the properties upon updates.

Hello! My team experienced this, and we resolved it by adding the "desired_capacity" and "desired_capacity_type" properties to the Terraform "aws_autoscaling_group" resource, then modifying the resource for the lifecycle tags: ignore_changes = [desired_capacity, desired_capacity_type]. Basically adding the properties, which would be used upon resource creation, then ignoring the properties upon updates.

This worked for me thanks!